European

Competitiveness
Report 2013
TOWARDS
KNOWLEDGE DRIVEN
REINDUSTRIALISATION

Enterprise
andIndustry
European Competitiveness Report

2013

TOWARDS KNOWLEDGE-DRIVEN
REINDUSTRIALISATION

Commission Staff Working Document

SWD(2013)347 final
Cover picture: Fotolia_663404_Orlando Florin Rosu X

This publication is prepared by Directorate-General for Enterprise and Industry, Unit A5 "Economic Analysis and
Impact Assessment", under the overall supervision of Daniel Calleja Crespo, Director-General and Didier
Herbert, Acting Director of the Directorate for Industrial Policy and Economic Analysis.

The report partly draws on background studies prepared by a consortium led by WIFO (see back of report for a
full list of background studies). It was put together by a team consisting of Mats Marcusson (Chapter 1), Neil
Kay (Chapter 2), Anastasios Saraidaris (Chapter 3), Mats Marcusson and Marshall Hsia (Chapter 4), Tomas
Brnnstrm (Chapter 5), Luigi Cipriani (Statistical annex and support), Marieta Todorova, Elizabeth Catzel and
Patricia Carbajosa-Dubourdieu (administrative support). The team was led by Tomas Brnnstrm, Project
Manager, and Konstantin Pashev, Head of Unit.

Comments and suggestions by many colleagues from several Directorates-General (Enterprise and Industry;
Communications Networks, Content and Technology; Research and Innovation; Economic and Financial Affairs;
Health and Consumers; Trade; Competition; Enlargement; Environment; Climate Action; Joint Research Centre)
are gratefully acknowledged.

The report and related materials can be downloaded at:

http://ec.europa.eu/enterprise/policies/industrial-competitiveness/competitiveness-analysis/index_en.htm

For further information contact us

By e-mail to: entr-econ-analysis-and-ia@ec.europa.eu
Or by mail to: Directorate-General for Enterprise and Industry
Unit A5 - Economic Analysis and Impact Assessment Unit
BREY 06/181
European Commission
B - 1049 Brussels (Belgium)

This publication is financed under the Competitiveness and Innovation Framework Programme (CIP) which aims
to encourage the competitiveness of European enterprises.

More information on the European Union is available on the Internet (http://europa.eu).

Cataloguing data can be found at the end of this publication.

Luxembourg: Publications Office of the European Union, 2013

ISBN
ISSN
doi
European Union, 2013
Reproduction is authorised, provided the source is acknowledged, save where otherwise stated.

For use/reproduction of third-party copyright material specified as such permission must be obtained from the
copyright holder(s).

Legal notice:
Neither the European Commission nor any person acting on its behalf may be held responsible for the use to
which information contained in this publication may be put, nor for any errors which may appear despite
careful preparation and checking. The publication does not necessarily reflect the view or the position of the
European Union. Luxembourg, Office for Official Publications of the European Union, 2013.

Printed in Luxembourg
 Foreword

Daniel Calleja Crespo

Director General, DG Enterprise and Industry

Dear Reader

This years European Competitiveness Report coincides with the first preliminary signs that Europe is finally
joining the rest of the world on the path to steady recovery, leaving the worst years of the crisis behind. The
report is dedicated to the role of manufacturing in making this growth path irreversible and sustainable in the
long term.
Despite its declining share of EU GDP,
manufacturing is widely acknowledged as the
engine of the modern economy. This is due to its
lead contribution to overall productivity; to its
input to research and innovation, which is four
times higher than its input to GDP; and to its
multiplier effects on growth in the rest of the
economy.
The volume and power of this engine, however, is
declining vis--vis the overall weight it carries,
including private and public service sectors.
The declining share of manufacturing in EU GDP
is not a new phenomenon. Part of the explanation
is related to long-term structural change. With the
growth in personal incomes, advanced economies tend to consume more services than manufactured goods.
Secondly, services have been relatively shielded from price competition as they are less mobile than
manufactures. This keeps their relative prices higher, with more attractive and often faster returns than
manufacturing. Thirdly, some services have low price elasticities: households cannot react to high prices by
reducing consumption (e.g. health, education or legal services); in the same way firms cannot freely adjust their
consumption of compliance-related services (accounting, audit, reporting, conformity assessments, legal
services, other professional business services, information to consumers and authorities, etc.) and other business
related services. Finally, with the increasing specialization of manufacturing and the high share of small
businesses, a growing share of the above mentioned services are bought in the market rather than produced in-
house, thus pushing further the structural shift to a service-dominated economy. These trends have been
observed and empirically established for quite some time.
What is new however is that in the last decade the shift away from manufacturing in Europe has accelerated,
reaching a critical threshold below which the sustainability of the European economic and social model might be
at risk. This is partly due to the financial and construction bubbles prior to the crisis and partly because of the
faster decline of manufacturing relative to services during the crisis. But the report also shows that a large part of
this phenomenon stems from growing competitive pressures from emerging industrial powerhouses, and this is
not only in the low-tech homogenous products where competition is mainly on price. If the EU economy is to
return to the path of sustainable and inclusive growth and find solutions to the pressing societal challenges of the
21st century, we need a larger and more powerful manufacturing engine to take us there.
The crisis has left little doubt in this regard. The report documents that the recovery has been driven mainly by
exports of manufactures, benefiting from the EUs preserved and upgraded comparative advantages in high-end
products. Despite the crisis, the EU has a comparative advantage on the world markets in about two thirds of the
manufacturing sectors, which account for of EU manufacturing value added. The report also finds that these
comparative advantages are concentrated in products with a high degree of sophistication and knowledge
intensity. Additional evidence of the sophistication and resilience of the EU industrial base is its higher domestic
content of exports relative to its competitors. And this is not because EU enterprises are less integrated in the
global value chains: EU content in US, Chinese and Japanese exports is higher than other foreign content in
these countries' exports. Taken together, this evidence shows that the EU can rely more on local supply chains
for high-tech inputs than our major competitors. These are strengths and advantages which provide a firm ground
for further gains and upgrades of EU industrial competitiveness, and for increasing the volume and power of the
manufacturing engine of the EU economy.
The report, however, reveals a number of challenges which call for an urgent and well-targeted policy response.
In the high tech sectors, the EU has comparative advantage in pharmaceuticals but lags behind in the rest of this
broad category (computers, electronics, optical equipment). Even in the medium high-tech sectors, EU
comparative advantage is lower than for the US and Japan. More importantly, China and the other emerging
industrial powerhouses are quickly gaining ground in the knowledge intensive sectors and rather than merely
assembling high-technology products they are now producing them.
The report looks at the EU-US productivity gap and finds that after a short period of narrowing prior to the crisis,
it is widening again. A decomposition of US higher productivity gains vis--vis the EU shows that they are
accounted for by a higher contribution of investment in ICT and by higher total factor productivity gains. The
report looks as well at the contribution of the technical efficiency gap to EU productivity underperformance and
derives the relevant implications for industrial policy. Business expenditures on R&D in the EU remain
considerably below the US. More importantly, the report provides evidence that this is not due to a difference in
industrial structure, but to a lower level of R&D spending across all sectors. Finally, the report sheds light on the
slower market uptake of research results in the EU.
These are important issues which mark the transition in our policy agenda from crisis management to smarter,
longer-term and more coherent governance of EU industrial competitiveness, which will bring us back on the
road of sustainable and inclusive productivity growth. They will be at the centre of the EU industrial policy
debate in the run-up to the elections of the new European Parliament in May 2014. Your voice, the voice of
private and public sector experts, academia, entrepreneurs, consumers and employees in this debate is more
important than ever. Therefore I would invite you to participate in our discussions (including online), and share
with us your thoughts and ideas about the future of European manufacturing. I hope that the empirical evidence
and analysis presented here provide interesting and inspiring pointers in this debate.

Daniel Calleja Crespo

 Contents
EXECUTIVE SUMMARY.................................................................................................................................... 3
Chapter 1. The competitive performance of EU manufacturing .................................................................... 15
1.1. Delayed recovery................................................................................................................................ 15
1.2. EU industries performances on world markets ................................................................................. 17
1.3. Drivers of sectoral competitiveness.................................................................................................... 22
Chapter 2. Structural Change ............................................................................................................................ 41
2.1. Broad Trends in Structural Change .................................................................................................... 41
2.2. Productivity improvements and changes in demand as drivers of structural change ......................... 45
2.2.1. Interaction of supply and demand factors........................................................................................... 45
2.3. The expansion of the Services Sector ................................................................................................. 47
2.3.1. Interaction of manufacturing and services.......................................................................................... 50
2.4. Heterogeneity of structural change in Europe .................................................................................... 52
2.5. International trade and specialisation as drivers of differences in the economic structure ................. 53
2.5.1. The role of institutions in structural change ....................................................................................... 58
2.6. Summary and policy implications ...................................................................................................... 61
Chapter 3. Reducing productivity and efficiency gaps: the role of knowledge assets, absorptive
capacity and institutions.................................................................................................................. 69
3.1. Growth accounting and the effect of the crisis at country and sector level ........................................ 70
3.1.1. Economic performance of the EU and other major economies: overview of aggregate output and
productivity trends.............................................................................................................................. 70
3.1.2. Growth accounting analysis: Sources of productivity growth at aggregate level ............................... 72
3.1.3. Productivity developments in the EU and the United States: a sectoral perspective .......................... 73
3.2. The role of knowledge transfer, absorptive capacity and institutions for productivity growth .......... 76
3.2.1. Empirical model and estimation ......................................................................................................... 77
3.3. Efficiency analysis at the industry level ............................................................................................. 81
3.3.1. Reducing efficiency gaps: discussion of the main determinants ........................................................ 84
3.4. EU productivity performance at the firm level: evidence from EU-EFIGE survey on
manufacturing firms ........................................................................................................................... 87
3.4.1. Analysis of the productivity effect of the crisis: econometric evidence ............................................. 90
3.5. Summary and policy implications ...................................................................................................... 93
Chapter 4. A manufacturing imperative in the EU: the role of industrial policy ..................................... 105
4.1. The manufacturing imperative in a European Context ..................................................................... 106
4.2. The main source of innovation and technological progress ............................................................. 106
4.3. Increased linkages between manufacturing and services ................................................................. 107
4.4. The carrier function of manufactures ............................................................................................. 108
4.5. Productivity growth .......................................................................................................................... 108
4.6. Does manufacturing offer higher wages in Europe? ........................................................................ 109
4.7. Structural Change in the EU Economy............................................................................................. 109

1
 4.8. Trends within EU manufacturing ..................................................................................................... 111
4.9. External competitiveness .................................................................................................................. 112
4.10. R&D as means to meet competition ................................................................................................. 115
4.11. Industrial policy measures in the European Union ........................................................................... 116
4.12. Industrial policies at the Union level and by Member States ........................................................... 116
4.13. Quantitative assessment of state aid and export orientated manufacturing ...................................... 119
4.14. Company-level analysis of commercialisation performance ............................................................ 121
4.15. Effects of public funding for firms' R&D ......................................................................................... 126
4.16. Summary and policy implications .................................................................................................... 128
Chapter 5. EU production and trade based on key enabling technologies................................................... 147
5.1. Technology positions and market potential ...................................................................................... 147
5.1.1. Introduction ...................................................................................................................................... 147
5.1.2. Approach .......................................................................................................................................... 148
5.1.3. Industrial biotechnology ................................................................................................................... 148
5.1.4. Photonics .......................................................................................................................................... 149
5.1.5. Micro-/nanoelectronics ..................................................................................................................... 150
5.1.6. Advanced materials .......................................................................................................................... 152
5.1.7. Nanotechnology ............................................................................................................................... 152
5.1.8. Advanced manufacturing technologies ............................................................................................ 153
5.1.9. North American decline, East Asian rise.......................................................................................... 154
5.2. The position of Europe in the production and trade of KETs-related products ................................ 155
5.2.1. Introduction ...................................................................................................................................... 155
5.2.2. Technology content of products related to key enabling technologies ............................................. 155
5.2.3. Type of competition and competitive advantages in international trade .......................................... 158
5.2.4. Link between patenting and technology content of products related to key enabling technologies . 161
5.3. Value chain analysis of promising KETs-based products ................................................................ 162
5.3.1. Introduction ...................................................................................................................................... 162
5.3.2. Selection of products ........................................................................................................................ 162
5.4. Value chain analysis of lipase enzymes ........................................................................................... 164
5.4.1. Value chain decomposition .............................................................................................................. 164
5.4.2. EU activity along the value chain ..................................................................................................... 164
5.4.3. EU position in the value chain of lipase enzymes ............................................................................ 167
5.5. Value chain analysis of the accelerometer ....................................................................................... 167
5.5.1. Value chain decomposition .............................................................................................................. 167
5.5.2. The EU position in the value chain of the accelerometer ................................................................. 170
5.6. Conclusions and policy implications ................................................................................................ 170
Chapter 6. Statistical annex .............................................................................................................................. 177
6.1. Sectoral competitiveness indicators ................................................................................................. 177
6.1.1. Explanatory notes ............................................................................................................................. 177

2
 EXECUTIVE SUMMARY
MAIN FINDINGS OF THIS REPORT:
Although the weight of manufacturing in the EU economy is decreasing in favour of
services, manufacturing is increasingly seen as a pivotal sector. However, critical mass
in the form of a minimum production base is needed. Industrial policy supporting
innovation and external competitiveness can play a role to reverse the declining trend.
To this end, EU industrial policy needs to steer structural change towards higher
productivity in manufacturing and better positioning of EU enterprises in the global
value chain based on comparative advantages in knowledge and technology intensive
products and services.
This is a must and a challenge for two reasons. First, the EU is lagging behind in
productivity gains relative to emerging industrial powerhouses and some of its major
competitors. The EU-US productivity gap, for instance, is growing wider again after
years of narrowing. It is linked to a production efficiency gap caused by regulations,
lower investment in ICT and intangible assets. In some sectors there is also a
commercialisation of research gap between the EU and the US. Policies targeting not
only creation of new technologies, but also knowledge diffusion through measures to
stimulate the supply of skills on the one hand, and demand for R&D on the other can
help bridge such gaps.
Second, structural change is slow, path-dependent and needs to build on existing
strengths, but can be stimulated by having the right institutional framework in place,
covering education, research, technology and innovation policies but focusing also on
the general quality of governance.
On the positive side, the report documents that the existing strengths of EU
manufacturing are substantial. The revealed comparative advantage of EU
manufacturing is linked to complex and high-quality product segments. By gradually
increasing the complexity of their products, EU manufacturing industries managed to
maintain their competitive position in 2009 compared with 1995. Moreover, EU
manufactured exports have less embedded foreign value added than exports by third
countries such as China, South Korea, Japan and USA.
The EU is a major producer of new knowledge in key enabling technologies. Its
products based on industrial biotechnology or advanced materials have higher
technology content than competing North American or East Asian products. Apart
from advanced manufacturing technologies, EU products based on key enabling
technologies are mature and need to compete on price. Adding more innovative and
complex products to the product portfolio will help manufacturers move up the value
chain.
THE COMPETITIVE PERFORMANCE OF EU MANUFACTURING
This years edition of the European Competitiveness Report uses a number of traditional and
advanced indicators of industrial competitiveness to provide insights into the strengths and
weaknesses of EU manufacturing and draw implications for EU industrial policy. It shows
that the EU has comparative advantages in most manufacturing sectors (15 out of 23)

3
accounting for about three quarters of EU manufacturing output. They include vital high-tech
and medium-high-tech sectors such as pharmaceuticals, chemicals, vehicles, machinery, and
other transport equipment (which includes aerospace).
EU MANUFACTURING VALUE CHAINS CAN SUPPLY HIGH-TECH
INTERMEDIATES FROM THE HOME MARKET
Furthermore, the report evaluates industrial competitiveness by looking at trade in value
added to analyse the place of EU manufacturing in the global supply chains. The domestic
and foreign content of a country's exports provide information on whether that country
develops or merely assembles high-technology products. Analysis of manufacturing exports
from China, the EU, Japan, South Korea and the US from 1995 to 2009 shows that foreign
value added embedded in EU manufacturing exports the part of value added coming from
inputs imported from other parts of the world is lower than for other countries. Conversely,
EU added value in the exports of emerging industrial powerhouses increased more than that
from other parts of the world. Between 1995 and 2009, when Chinese exports increased
dramatically, EU industries managed to increase their value added content in Chinese
manufacturing exports more than industries from other parts of the world. Japanese, South
Korean and US value added content shares of Chinese manufacturing gross exports decreased
during the same period. Summing up, the report finds that the EU has a higher share of
domestic content of exports than established and emerging industrial competitors, while at the
same time has a higher share of its intermediates in other countries' exports. This is evidence
of a strong industrial base which allows EU enterprises to source most of their high-tech
inputs (goods and services) domestically, while also supplying them to the rest of the world.
EU MANUFACTURING EXPORTS HAVE HIGHER DEGREE OF COMPLEXITY
A further evidence of the industrial strengths of the EU is the analysis of the sophistication
(knowledge intensity) of EU exports of products with comparative advantages. This is an
advanced indicator of non-cost competitiveness which shows that manufacturing industries in
the EU have a higher degree of complexity. The report documents that EU exports have
preserved their advantages thanks to developing sophisticated, knowledge-intensive products
to address the cost advantages of emerging industrial powers. By gradually increasing the
complexity of their products from 1995 to 2010, EU manufacturing industries managed to
maintain their competitive position. By contrast, products from BRIC countries (Brazil,
Russia, India, China) underwent major changes in the same period goods produced by firms
in wood industries, radio, TV and communication equipment industries, medical, precision
and optical instruments industries, and furniture industries in BRIC countries have
considerably improved in terms of their average complexity but the majority of industries in
BRIC countries still produce less complex products than EU industries. As a consequence, in
2010 the EU exported around 67% of products with revealed comparative advantage, while
China had comparative advantage in 54% of its exports, the US in 43% of its products, and
Japan in 24%.
A ROLE FOR INDUSTRIAL POLICY
The report however documents trends and developments which call for urgent and well-
targeted industrial policy measures to build on the identified strengths and upgrade the
competitiveness of EU manufacturing.
Of the 15 sectors with comparative advantages mentioned above, about two-thirds are in the
low-tech and medium-low tech manufacturing groups. On a positive note though, even in
those sectors EU competitiveness is based on high-end innovative products.

4
In the high-tech sectors, the EU has comparative advantages in pharmaceuticals but lags
behind in computers, electronics and optical products, while in the medium-high tech
industries its comparative advantages are lower than in the US and Japan. China and other
emerging industrial powerhouses are also quickly gaining ground in the international
competition in the knowledge-intensive sectors, successfully upgrading their exports from
assembly to developing high-tech products and knowledge-intensive services.
With a view to identifying the drivers of non-cost competitiveness, the report looks at
indicators of skills and investment in physical capital and intangibles to draw the relevant
policy implications for guiding EU manufacturing to knowledge-based productivity gains.
US private spending on R&D (as a share of GDP) is almost 1.5 times that of the EU (2.7% in
the US; 1.85% in the EU). A sector breakdown indicates that this is not a result of differences
in industrial structures or US specialisation in knowledge-intensive sectors, but due to an
overall underperformance of EU sectors in terms of R&D investment across all sectors. The
output of research is new products, new technologies, new materials and processes. A rough
indicator of this output is patents. The report documents that in a number of high and
medium-high technology industries (such as pharmaceuticals, optical equipment, electrical
equipment, medical and surgical equipment, telecom and office equipment, radio and TV and
accumulators and batteries), the EU is lagging behind in terms of patenting. As the RCA
(revealed comparative advantage) indicators show, EU export performance depends crucially
on some of these sectors. It may be hard to preserve current EU comparative advantages in
these industries if the EU loses its technology lead (as indicated by patent data). Another
problem is that the transmission of research results from the laboratory to the market which
seems to be more difficult in the EU than for its major competitors.
The implications for EU industrial policy of its export complexity is that targeting only high-
tech sectors might be less rewarding than increasing the share of knowledge-intensive
products in all tradable sectors, including medium-low tech and medium-high tech sectors.
Moreover, some of the labour-intensive sectors with lower knowledge intensities may be
better suited to tackle the EU's unemployment challenges than the high-tech sectors. About
40% of EU manufacturing employment is in low-tech sectors. Therefore the policy priority
attached to key enabling technologies which lead to new materials and products in all
manufacturing sectors has a strong potential to upgrade EU competitiveness not only in the
high-tech sectors but also in the traditional industries.
LONG-TERM DYNAMICS OF STRUCTURAL CHANGE
Almost all countries follow a broadly similar pattern of structural change. As economic
development gets under way, the share of agriculture in national employment and value added
falls, and there is a rapid increase in the share of manufacturing and services. The resource
reallocation process associated with structural change shifts economic activities from
agriculture to industry and services.
DRIVERS AND IMPACT OF STRUCTURAL CHANGE
There are two central, possibly complementary, theories of the observed patterns of structural
change. The first, supply-side, explanation highlights the differential patterns of technical
change between sectors. Here, structural change can be viewed as a consequence of
differential productivity growth rates across agriculture, industry and services, where
technological progress is the main driving mechanism behind productivity growth. The
second, demand-side, theory relates structural change to different income elasticities of
demand between products and services of different sectors. These different elasticities provide
a sorting mechanism on the development of sectors.

5
The increasing contribution of the service industry, at the expense of manufacturing, can also
be partly explained by an increasing service content of manufacturing final output. This
content reflects the total value of the services required for the development, production and
marketing of a modern manufacturing product. The service content of manufacturing has been
growing in the EU and elsewhere in the world. Currently about a third of the price of a
manufacturing product in the EU is associated with integral services. Whilst manufacturing
products too are used for producing services, the manufacturing content of services produced
in the EU is only around 10 per cent.
The gradual rise in services and reduction in the manufacturing share of valued added do not
mean that manufacturing can be ignored. It is still seen as a pivotal, though heterogeneous,
sector with important production and demand linkages that play a significant role in the
process of economic development.
The analysis in Chapter 2 confirms that the structural change across economies produces
more diverse country profiles in manufacturing sectors than in services sectors. Wider
tradability of manufactured goods leads to more variability.
Analysis also confirms that structural change is gradual and path dependent based on the
specific capacities and capabilities of individual economies, which are important determinants
of sector growth.
Structural change can generate both positive and negative contributions to aggregate
productivity growth. On average, structural change appears to have only a weak impact on
aggregate growth over short time periods.
THE ROLE OF INSTITUTIONS IN STRUCTURAL CHANGE
Institutions can positively affect structural change in a number of ways. For example,
differences in the patterns of technology diffusion are considered to account for a sizable part
of the divergence in incomes between rich and poor countries. Educational attainment can
also be linked to product specialization patterns, with a positive correlation between high
knowledge intensity and product complexity. Microeconomic evidence also suggests that
credit market imperfections are important sources of differences in productivity across
countries. Market frictions can also hinder structural change due to the existence of
regulations and administrative burdens that inhibit the reallocation of resources across sectors
and firms. Many factors can be identified such as certain types of taxes, labour market
regulation, size-dependent policies or trade barriers in addition to regulations and costs of
doing business in the formal sector.
STRUCTURAL CHANGE: POLICY IMPLICATIONS
Structural change is thus dependent on progressive policies and institutions that allow an
efficient allocation of resources within economies. Policies and institutions that hinder such
reallocation are a source of inefficiency and impede economic development. Policies to foster
structural adjustments should be conceived in a broad way and span such different areas such
as education policies, research, technology, and innovation policies, but also focus on the
general quality of governance.
International trade is an important determinant of the development of sectoral shares in
countries. The successful catch-up stories of Germany in 19th century, and Japan and South
Korea in the 20th century, cannot be explained without taking into account international trade,
comparative advantage in tradable goods and specific competencies and capabilities in the
production of new and high-value added products. Here it is important to acknowledge that
structural change that shapes economic development of countries is highly path-dependent

6
and cumulative. Any change is rooted in present knowledge bases and constrained by existing
specialisation patterns. Complementary capabilities need to be built up. Thus policies to
support structural change should always start by taking into account the existing production
structures of countries and regions, as well as the knowledge base of supporting institutions.
Countries seeking to shift their industrial production up the technology ladder are likely to
also need to increase and improve non-government services, such as education and business
services.
The centrality of institutions and policies in the process of structural change leads to a view
that the general quality of institutions is important to structural change. Policies that foster
structural adjustments should therefore be conceived in a broad way and cover such different
areas as education, research, technology and innovation policies, while also focusing on the
general quality of governance.
DIFFERENCES IN PRODUCTIVITY GROWTH BETWEEN THE EU AND THE US:
EVIDENCE FROM A GROWTH ACCOUNTING ANALYSIS
In the late 1990s and early 2000s, the US productivity lead resulted from a first mover
advantage in ICT as is illustrated by the relevant growth accounting analysis. The EU
experience reveals that movements in Total Factor Productivity (TFP) have followed US
productivity developments, but with a time lag.
Prior to the financial and economic crisis of 2007-2008, the debate on the European
productivity slowdown focused on the slower adoption of new technology as the main reason
behind the EUs relative productivity under-performance. Moreover, industry-based studies
revealed that the US productivity advantage was found in a few market services sectors,
mainly trade, finance and business services.
The initial hypothesis was that the EU was lagging behind the US merely in the adoption of
ICT but would eventually benefit from the same productivity gains. Chapter 3 reveals that
high investments alone are not sufficient to boost economic growth and to guarantee a better
productivity performance.
The EU is not only still lagging behind the US, but the productivity growth gap has recently
increased.
THE CHANGES IN THE EU-US PRODUCTIVITY GAP FROM A SECTORAL
PERSPECTIVE
The European failure to match the US acceleration in output and productivity between 1995
and 2004 has largely been attributed to developments in market services. The analysis reveals
that the sector which contributed most to amplify the US productivity advantage was
wholesale and retail trade, due to its strong productivity performance and its relatively large
share in the economy. Other service sectors with sizeable contributions included professional,
scientific, technical, administrative and support services, plus finance and insurance activities.
Throughout the period 2004-2007, two factors contributed to the reduction of the EU-US
productivity gap, the acceleration of productivity in most EU manufacturing industries, and a
robust performance of many EU services sectors.
In the US wholesale and retail sector, labour productivity slowed significantly between 2004
and 2007 compared to the exceptional performance observed in the previous period. On the
other hand, the EU performance in the same sector improved substantially, reaching 3%
productivity growth, nearly double the rate achieved in previous periods. In most EU services,
labour productivity improved, particularly in professional, scientific, technical activities, and
community, social and personal services. Overall, between 2004 and 2007, those sectors that

7
contributed to narrowing the EU productivity gap relative to the US were the same ones
which had caused EU productivity to stagnate in the previous decade.
During the financial crisis (2007-2010), labour productivity stalled in the EU, while in the US
it continued to improve. The majority of manufacturing sectors in the EU experienced a fall in
productivity levels, probably reflecting a higher exposure to global demand fluctuations than
the services sectors. Manufacturing productivity as a whole, decreased by more than 1%
annually, with chemicals down by more than 4%. In the US, manufacturing productivity grew
by over 4% a year during 2007-2010. One of the few sectors to experience a worsening in
productivity levels was chemicals. The majority of services activities though, experienced
robust growth, particularly telecommunications, finance, insurance, IT and information
services.
The sectors where the US productivity advantage increased are electrical and optical
equipment and the majority of manufacturing sectors, as well as construction, and
telecommunications. On the other hand, service sectors such as financial activities, business
services, accommodation, food and some public services are among the EU sectors that
helped narrow the gap during the most recent years. In the EU manufacturing sector, TFP
was the main driver of the output growth up to 2007 but it was also the main cause of the
declining productivity thereafter. In services, the picture is more heterogeneous. In the US,
substantial TFP gains explain the productivity acceleration relative to the EU in the late
1990s. The finance and insurance sector in the US also experienced considerable ICT-driven
productivity growth in the late 1990s and early 2000s.
TFP improvements in the EU wholesale and retail sector took place with some years delay
and contributed to closing the gap in the period just prior to the financial crisis. Since the
crisis, the EU finance and insurance sector has also shown a considerably better performance
than in the US.
THE ROLE OF KNOWLEDGE TRANSFER, ABSORPTIVE CAPACITY AND
INSTITUTIONS FOR PRODUCTIVITY GROWTH
In the EU, too little has been invested in the skills and organisational changes necessary to
reap the benefits of ICT technologies. Lower investments in intangible assets (R&D, human
capital, etc.) are likely to explain a portion of the US-EU productivity gap as these factors
affect a countrys absorptive capacity, i.e. its ability to take advantage of technology
developed elsewhere (international technology transfers). Given that the bulk of technological
innovations is concentrated in a few leading countries, improvements in the absorptive
capacity will be needed in order to assimilate foreign technologies.
DETERMINING THE EU-US EFFICIENCY GAP
Understanding why industries vary in their ability to use resources effectively, and identifying
suitable policies to improve efficiency performance, requires the analysis to look into factors
that cause industries to lose productivity and hence widen efficiency gaps.
The empirical results show that ICT plays a key role in reducing inefficiencies in the use of
resources. In addition, more upstream regulation significantly increases the efficiency gap. In
other words, administrative restrictions imposed on service market competition have
widespread negative effects on production efficiency.
These results provide strong support for the hypothesis that a more competitive business
environment reduces the efficiency gap. More flexible product market regulations, largely
concentrated in key service-providing industries, are likely to raise efficiency levels across the

8
whole economy. Regulatory changes in the labour market should also be tailored to restore
the necessary balance between regular and temporary workers.
A few market service sectors were the main cause of the EU productivity disadvantage
compared to the US during the emergence of ICT technologies (in the late 1990s). However,
in the years leading up to the financial crisis, the EU experienced strong ICT-related labour
productivity growth in these sectors, mirroring earlier developments in the US, and thereby
reaching the US productivity levels. Since the crisis, the EU-US productivity gap has widened
again.
Chapter 3 examines two main channels for raising productivity growth potential and closing
the gap with the technology leaders. The first is the role of absorptive capacity and
knowledge-base (intangible) assets (R&D and human capital) in activating international
technology transfers. This mechanism has been found in the literature to be highly conducive
to TFP growth through spillovers. However, its growth-enhancing effect is heterogeneous; the
ability to accommodate the inflow of new technological knowledge by re-allocating factors or
expanding new product lines is required.
The second channel is via production efficiency as a possible factor behind the widening
productivity gaps between the EU and the US. There is evidence that technical efficiency is
significantly higher in countries with less restrictive product market regulations or
employment protection laws. Investments in ICT assets, on the other hand, help in reducing
the gap with the most efficient country and/or industry.
Intangible assets (e.g. R&D, human capital, organizational change, etc.) are important sources
of TFP growth and sustained long-run competitiveness. In this context, initiatives which
stimulate investments in these areas may be particularly useful.
Similar measures may also be put in place to increase the number of qualified staff per firm.
These measures could facilitate the hiring of highly qualified workers or promote workforce
training. Other policies could be directed towards enhancing inputs such as ICT which can
assist in the reorganisation of production. Specific ICT applications, such as enterprise
software systems, increase productivity at firm level. These measures would also be viable for
smaller firms which do not always have the necessary financial and human resources to
undertake R&D activities and for that reason, seek alternative ways of increasing their
competitiveness.
As regards the key role of financial resources on productivity performance, policies which
increase access to finance for SMEs are needed. The above measures should be conceived and
applied within an appropriate regulatory framework that will safeguard the stability of market
regulation and facilitate the reduction of productivity and efficiency gaps.
A MANUFACTURING IMPERATIVE IN THE EU DECLINING
MANUFACTURING SHARES OF VALUE-ADDED AND EMPLOYMENT
There are at least two well-documented reasons for the declining share of manufacturing
value-added in GDP.
Firstly, the higher productivity growth in manufacturing implies that in the longer term prices
of manufactures will decline relative to services, leading to a lower share of manufactures in
value-added in nominal terms. Therefore, a declining value-added share of the manufacturing
sector per se is not a reason for concern, but rather the logical consequence of a European
manufacturing sector that is constantly becoming more efficient.
Secondly, demand structures characterised by low price elasticities of demand and high
income elasticities for some services like education, tourism, health and cultural activities

9
change the composition of demand as income increases. These elasticity differentials
(discussed above in the context of structural change) compound the effect of the declining
relative importance of manufacturing in value-added terms.
These two factors driving the downward trend of a manufacturing sector in relative terms can
be countered by increased external competitiveness of the manufacturing sector and industrial
policies working towards this end, as set out in Chapter 4.
ARGUMENTS FOR THE MANUFACTURING IMPERATIVE
Fears have been raised that the declining manufacturing share of GDP entails loss of
manufacturing capabilities which, once lost, are hard to recover. Manufacturing capabilities
specific to particular industries even if they are low-technology industries may at a later
stage become important inputs for fast-growing new products.
There are at least three arguments for a critical size of the manufacturing base:
Manufacturing still accounts for a major part of the innovation effort in advanced
economies and this translates into above-average contributions to overall productivity
growth and thus to real income growth.
There are very important backward linkages from manufacturing to services which
provide important inputs for manufacturing (in particular business services).
Manufacturing has a carrier function for services which might otherwise be considered
to have limited tradability. This operates through international competitive pressure and
has an added stimulus effect for innovation and qualitative upgrading for service
activities. Another linkage is increased product bundling of production and service
activities in advanced manufacturing activities.
Lastly, and related to the first argument, is the higher productivity growth in
manufacturing which is important because the sector of origin of productivity growth
may not be the sector which benefits most from the productivity growth.
THE EXTERNAL COMPETITIVENESS OF EU MANUFACTURING
There are also structural changes within the manufacturing sector, as explained in Chapter 2
above. These go in the direction of a mild but persistent shift towards more technology-
intensive industries (chemicals, machinery, electrical equipment and transport equipment)
which also tend to be less labour-intensive.
This mild trend towards advanced manufacturing industries reflects international
specialisation patterns of EU Member States because in general technology-intensive
industries offer more possibilities for building comparative advantages by product
differentiation and quality aspects. At the same time, low-technology-intensive industries still
accounted for almost 40% of EU manufacturing employment in 2009.
Traditionally, EU manufacturing has faced competition in more technology-intensive
segments from producers in Japan, Korea or the US. However, over time competition from
producers in BRIC countries is gradually changing and increasing.
Given the structural upgrading in emerging economies, competitive pressures from these
countries are not limited to low-technology-intensive industries but are also felt in advanced
manufacturing industries. Brazil, India and China all increased considerably their market
shares in global value added exports of manufactures over a period of 15 years. It is especially
the outstanding performance of Chinese producers, whose market share quadrupled between
1995 and 2011, which drove this change.

10
All in all, EU manufacturing seems to have managed to defend its positions on world markets.
In general, the EU manufacturing sector is seen as rather well-diversified. Over time, the EU
manufacturing sector has succeeded in upgrading product quality by engaging in R&D&I.
The challenge is more demanding for low-tech and medium-low-tech industries for which this
may require a higher degree of specialisation and entering or creating niche markets. Existing
evidence suggests that many European firms follow such a premium strategy within their
respective industry. European firms typically operate in the top quality segments.
R&D in manufacturing is key to maintaining or expanding market shares for knowledge-
intensive goods. It is therefore worrying that EU manufacturing has a lower R&D intensity
per firm than the US and Japan. The R&D intensity in seven EU Member States, for which
data are available, is only 62% of that of the United States.
R&D and innovation are not the sole ingredients for a highly productive and internationally
competitive manufacturing sector. In order to differentiate products and charge higher price-
cost mark-ups, manufacturing firms depend increasingly on sophisticated services inputs.
The ever tighter inter-linkages between the manufacturing sector and the increasingly
dominant services sector in the EU economy work in two ways:
Increased use of services inputs and services embedded in manufacturing products can
increase the EU manufacturing sectors competitiveness
Through this increased interdependence, manufacturing can increase the tradability of
services.
INDUSTRIAL POLICY AND THE EXTERNAL COMPETITIVENESS OF EU
MANUFACTURING
Sectoral aid does not show any significant effects on extra-EU exports or value added per
capita for export oriented firms. On the other hand, internationalisation measures are
horizontal and have significant and positive effects on extra-EU exports.1
Other effective horizontal aid measures seem to be regional aid and aid for training of the
employees.
An industrial policy providing funds to different parts of the innovation process have positive
and significant effects on R&D intensities and patent application propensities for
manufacturing firms in the EU-15 and EU-12 irrespective of firm size and technological
intensity of the firm.
The effects on output of the innovation process, of the amount of the innovative sales or of
public funding differ according to the geographical location of the firm, its size and its
technological intensity. Public funding has positive and significant effects, in particular for
EU-15 firms. Further positive and significant effects are found for high-tech and medium-
high-tech and for SMEs.
These results suggest that there is potential to improve the targeting of public support and to
make it more effective. Especially in the EU-12, and irrespective of the actual objectives of
the support programmes, governments end up providing innovation support more often to
larger firms than to their smaller competitors. Given that small firms in particular face
considerable financial problems due partly to asymmetric information flows, they should be
the primary target of public funds.

1
For the purposes of this report, internationalisation measures mean horizontal measures aimed at supporting internationalisation of
commerce. (Export aid is generally prohibited under EU state aid rules).

11
The special targeting of grants is one way to improve the allocation of public funds to SMEs.
Other initiatives could include information campaigns about credits, cost-deductions and
subsidised loans for new entrepreneurs. As financial problems occur mainly in the
commercialisation phase, fostering venture capital investment would be another starting point.
RESEARCH COMMERCIALISATION
The EU is usually perceived as less effective at bringing research to the market when
compared to its main competitors such as the US, Japan, and South Korea. The relative
underperformance in research commercialisation in the EU has been attributed to a number of
factors including the absence of an entrepreneurial culture and a less developed venture
capital sector. The discussion about the main factors explaining the European innovation gap
is related to the so-called European innovation paradox which suggests that Europe does not
lag behind the US in terms of scientific excellence, but lacks the entrepreneurial capacity of
the US to effectively commercialise inventions and step thereby on an innovation-driven
growth path.
Analysis on the specific innovation-related factors and types of public funding on
commercialization performance focused on the commercialisation of R&D efforts at firm
level. In particular, the actual R&D performed internally and/or acquired from external
sources, the research collaboration activities with different players such as customers,
suppliers, public research institutions and other firms, and the firms use of particular types of
public funding for innovation were examined, using the Community Innovation Survey (CIS)
micro-data. Focusing on the commercialisation of R&D efforts, innovation output was
analysed in terms of innovative sales of companies.
The results of the analysis suggest that the impact of the R&D efforts on commercialization is
positive for both manufacturing and non-manufacturing firms. It is observed that the firms
which, in addition to their own R&D, also acquire R&D services externally tend to have
higher share of turnover from innovative products. This external acquisition of R&D results
can take place as a pure purchase of services, but also can be acquired in the framework of the
inter-firm R&D cooperation.
Concerning the different forms of R&D cooperation activities the results are mixed across
different groups and classes of firms. It can be seen that vertical cooperation (i.e. R&D
cooperation with suppliers and/or customers) is positively associated with higher
commercialization performance in firms coming from different size classes and different
technology intensity groups.
The effects of public funding on the commercialization performance of firms appear to be
positive in most classes and groups of firms considered. The relationship between the use of
local public R&D support and the commercialisation performance shows positive across all
different technology intensity domains. Public R&D support at the national level is positively
related to the share of innovative turnover. Firms appear generally to have higher
commercialisation performance when making use of EU-level public R&D support, with a
consistently strong and positive effect of public funding being found especially for firms in
medium-high and high-tech industries.
Bringing the most important findings together suggests a number of conclusions regarding the
general patterns of innovation and commercialisation performance of European firms. When
observing the behaviour of individual firms, the link between the R&D effort and the
commercialisation performance is rather pronounced and a positive relationship is observed in
number of cases, not only the R&D itself, but also its origin and the patterns of R&D
cooperation among firms play a role.

12
In particular, the results suggest that local R&D support does positively affect firm
commercialisation performance in all technology intensity and size classes. The effects of
national and EU funding are positive and significant for all firms and manufacturing firms
only, but mixed results are found for smaller subsamples. Overall, public funding has
consistently positive effects on innovative sales for medium-high and high-tech sectors firms,
while this statement is true to a lesser extent for firms in lower tech industries.
PRODUCTS AND TRADE BASED ON KEY ENABLING TECHNOLOGIES
Every two to three decades, an innovative concept or material comes along with the potential
to bring about fundamental change throughout the economy. Silicon arrived in the 1940s,
paving the way for the ICT revolution. In the 1970s, gallium arsenide made lasers ubiquitous
in DVDs, CDs and modern telecommunications. The late 1980s witnessed the interconnection
of several existing computer networks to create the internet. In the 1990s there was gallium
nitrite, which revolutionised photonics and in particular solid state lighting. Right now the
world is exploring the potential benefits of graphene, isolated as recently as 2004 and
subsequently acknowledged by the 2010 Nobel Prize in physics, as well as by the European
Commission which recently launched a ten-year flagship programme with a budget of EUR 1
billion to develop graphene technology. Decades from now, with the benefit of hindsight,
people may look back at graphene as another game-changing discovery.
The stakes and potential gains are high. In 2011, the European Commissions first High-Level
Group on Key Enabling Technologies (KETs) presented its final report which estimated the
market for key enabling technologies to be worth USD 1,282 billion by 2015 (photonics
480 bn; micro and nanoelectronics 300 bn; advanced manufacturing systems 200 bn; advanced
materials 150 bn; industrial biotechnology 125 bn; nanotechnology 27 bn). As 2015
approaches, it is of course crucially important to ensure that EU manufacturing is ideally
placed to benefit as much as possible from this potentially huge and growing market. To that
end, it is not enough for the EU to produce state-of-the art research results in key enabling
technologies, there must also be mechanisms in place to bring those results to market in the
form of commercial products, and there must be demand for the products. This was one of the
key conclusions of the first High-Level Groups report. This report develops that theme by
assessing the position of the EU in the production of and international trade in certain
products based on key enabling technologies, including changes in EU competitiveness over
time. Chapter 5 goes on to examine the specialisation of Member States in the production of,
and trade in, products based on key enabling technologies.
EUROPE IS A MAJOR PRODUCER OF NEW KNOWLEDGE IN KEY ENABLING
TECHNOLOGIES
In terms of knowledge production, Europe appears to be doing well. Measured by its share of
the global number of patent applications in each of the six key enabling technologies, Europe
is maintaining a similar share as North America (US, Canada, Mexico) in most key enabling
technologies, while East Asian patent applicants tend to be more productive than their
European and North American counterparts. Unlike North American applicants though,
European patent applicants have lost little or no ground in recent years: Europes shares of
global patent applications are similar to those reported in the 2010 edition of this report (EC
2010), whereas North American applications have fallen back. It is also important to
underscore that in absolute terms, European patent applications are increasing from year to
year in most key enabling technologies.
BUT IS NOT ALWAYS IN A POSITION TO BENEFIT IN THE FORM OF
PRODUCTS

13
But knowledge production is not synonymous with job creation and growth. In order to turn
patents into marketable products based on key enabling technologies, manufacturers need to
be well positioned in terms of the technology content of their products and in relation to the
competition they face on the global market. Unit value analysis indicates that EU products
based on industrial biotechnology and advanced materials have a higher technology content
than North American or East Asian products in the same fields, while in advanced
manufacturing technologies for other KETs the technology content is similar to North
American but higher than East Asian products. In nanotechnology and micro- and
nanoelectronics on the other hand, EU products have a relatively low and generally
decreasing technology content.
The technology content should be seen against the backdrop of the competitive situation in
which EU manufacturers have to sell their products. The analysis presented in Chapter 6
suggests that in all key enabling technologies except advanced manufacturing, EU
manufacturers are predominantly up against price competition, and in three technologies
industrial biotechnology, nanotechnology and advanced materials they are able to compete
on price. This finding is new and runs counter to the generally held view that production costs
are too high in the EU to enable manufacturers to compete on price. In photonics and micro-
and nanoelectronics, where price competition prevails as well, EU manufacturers tend to have
little or no price advantage and therefore struggle to compete with North American and East
Asian manufacturers. This does not mean that EU manufacturers in those fields should exit
the market or that policies to strengthen competitiveness should not be pursued. It simply
reflects the fact that historically EU manufacturers have not had a price advantage on a market
where price competition prevails.
The only key enabling technology in which quality competition dominates is advanced
manufacturing technologies for other key enabling technologies, where EU manufacturers are
able to compete with North American and East Asian rivals thanks to the superior quality of
their products. The impact of the high technology content of EU products manifests itself in a
possibility to compete with high-quality products even if they are more expensive than
competing products made in North America or East Asia.
MOVING TO THE HIGHER END OF THE VALUE CHAIN
Having to compete mainly on price (in five of the six key enabling technologies) may not be
an attractive growth model for EU manufacturers in the long run. Given its considerable
knowledge production and the high technology content of EU products, a gradual shift away
from the current portfolio of predominantly mature products where firms compete more on
price than quality to more innovative and complex products could be an avenue to pursue. A
step in that direction could be to focus on more integrated products than today, possibly
combining more than one key enabling technology. Another idea could be to reinforce the
cross-fertilisation of new technology developments between key enabling technologies.

14
 Chapter 1.
THE COMPETITIVE PERFORMANCE OF EU
MANUFACTURING
This chapter overviews the competitive performance decades. The decline accelerated with the onset of the
of EU manufacturing vis--vis established and financial crisis (Figure 1.1).
emerging global competitors. EU industries compete
on the single market and on external markets by Figure 1.1. Double-dip of EU manufacturing production
selling their products either at a lower price or with a
higher quality than competitors.2 Their ability to
compete depends on a number of drivers. Some of
these drivers are necessary in order to compete with
prices while other drivers are more essential for the
development of products with characteristics and
qualities that differentiate the industries products
from those of their competitors.3 The analysis relies
on a number of traditional indicators of international
competitiveness, such as revealed comparative
advantages, labour productivity and unit labour costs,
but employs as well relatively novel indicators of Source: Own calculations using Eurostat data.
exports in value added and complexity and
exclusivity of exports. Even though the financial crisis had global
repercussions, other parts of the world have been
The chapter is organised as follows. The first section recovering faster. While the EU manufacturing hit the
presents a brief overview of the impacts of the trough and began a rebound earlier than US
recession on the manufacturing sector. The second manufacturing, since late 2011 the EU has been
section focuses on the export performance of EU lagging behind. The recovery in the two previous
industries on world markets. The third section recessions since 1990 was also faster in the US than
explains the export performance by analysing the the EU. Asia is also recovering faster than Europe.
drivers of EU price and non-price competitiveness. It South Korean manufacturing for instance reached its
looks at the dynamics of labour productivity and unit pre-crisis peak in less than 18 months.5 Similarly, the
labour costs (ULC), as well as patenting and initial rebound of Japan which was hardest hit by
innovation output. R&D and innovation indicators the financial crisis was impressive, but was
however are not readily available for all the Figure 1.2. EU recovery in comparative perspective
comparator economies.4 Therefore they are used
mainly for comparisons across EU industries and
Member States.

1.1. DELAYED RECOVERY

EU manufacturing output decline reached its trough
in the middle of 2009. Following a short-lived
recovery manufacturing industries fell back into a
double-dip recession at the end of 2011. Employment
in manufacturing has been steadily declining for

Source: Own calculations using Eurostat and OECD

2
manufacturing output data.
This is a simplification more accurately describing the
situation of firms selling just one product. Firms and industries
5
also have other means of competing. Examples of other means
to compete are combinations of goods with services,
combinations of goods that are complementary to each other,
establishment of distribution networks.
3
See European Commission (2010) for thorough analyses and
discussions of price and non-price factors.
4
Data for R&D expenditures, business expenditures on R&D
(BERD), are unfortunately published with a long delay. At the
time of drafting this report, a comprehensive data set for EU
Member States and OECD countries, is not available after
2008.
A sharp depreciation of the won by 31% from the first quarter
of 2008 may partly explain the 10% growth in exports in 2009.
Close relations with other Asian countries is another factor
accelerating Korea's recovery. Especially the Chinese stimulus
programme in 2009 contributed significantly as Korean exports
to China accounted for 87% of the increase of exports during
2009. Other factors explaining the rebound of Korean
manufacturing were the strong domestic demand growth,
including fiscal expansion, and a relatively limited impact of
the global financial crisis on Korean financial markets (OECD
2011)

15
interrupted by the 2011 Fukushima earthquake and
tsunami (Figure 1.2).
Figure 1.3. EU manufacturing recovery by member
state
Source: Own calculations using Eurostat manufacturing output
data as of March 2013.
Recovery has been much harder across the EU. While
Romania, Poland, Slovakia and the Baltic states have
already surpassed their pre-recession peak levels of
industrial output, most of the member states are still
below, with some of those in the south still close to
the trough or may have not even started their recovery
(Figure 1.3).
Figure 1.4. EU manufacturing recovery by sector
Source: Own calculations using Eurostat data. Developments are shown since the peak in EU aggregate manufacturing output in January
2008 to March 2013. HT, HMT, LT and LMT denote high-tech, medium-high-tech, low-tech and medium-low-tech manufacturing industries
(see the annex for definitions)
A sector breakdown shows that few industrial sectors
(among which pharmaceuticals and other transport
equipment) have recovered their pre-crisis level of
production (Figure 1.4). In principle, high-tech
manufacturing industries were less severely impacted.
Food, beverages and non-durable consumer goods
have fared relatively better than other industries since
the outbreak of the financial crisis. The reason is that
capital goods and intermediate goods industries are
16
more sensitive to business cycle fluctuations than
industries producing necessity goods and non-durable
consumer goods, demand for which is less sensitive
to variations in income.6 Some medium-high
technology industries produce capital and
intermediate goods, which is why they experienced
larger output decline.
Mining and construction were harder hit than total
manufacturing. There is however a considerable
variation within the aggregate mining and quarrying.
Metal ores and mining support services have had a
positive development since 2008. Some of this
development is due to a high demand from the world
market. On the other hand, some mining industries
have been in decline for a longer period of time
before the crisis. This is also true of some
manufacturing industries such as furniture, clothing
and textiles.
Since manufacturing was hit more severely than
services industries, the shares of manufacturing to
GDP fell in every Member State during the crisis.7
Figure 1.5 presents the shares of manufacturing in
GDP by country in 2012. The declining share of
manufacturing output and employment has been a
long-term trend driven by shift in domestic demand
due to growth in incomes on the one hand and lower
prices of manufactures due to higher productivity
6
See the discussion in European Commission (2009, 2011).
7
See for example European Commission (2013c) forthcoming
on the developments of services since 2008. Total services
declined by some 9% between the first quarters of 2008 and
2013 while the corresponding decline for total manufacturing
amounted to some 12%. Certain services industries providing
services with high income elasticities, for example
transportation services and package holidays, experienced
much stronger declines.
growth on the other (Nickell et al. 2008).8 Increasing
external demand for EU manufactured goods
however can counter this trend provided that EU
manufacturing industries compete successfully on
world markets.
Figure 1.5. Manufacturing shares of GDP in the EU
2012
Source: Own calculations using Eurostat data. Note: 2011 value
for Romania No data available for Bulgaria and Ireland.
1.2. EU INDUSTRIES PERFORMANCES ON WORLD
MARKETS
The recession which began in 2008 had a global
impact. While public and private debt problems have
constrained domestic demand in many Member States
Figure 1.6. Contribution to GDP growth in per cent 2011-2012
Source: AMECO database, Commission services.
and delayed a full recovery from the crisis, demand
for EU exports has risen. During 2010-2011 exports
contributed more to GDP growth than domestic
demand in most EU member states (Figure 1.6).
8 9
See chapter 2 for detailed analysis
17
Following the decline in trade during 2009, world
demand recovered faster than in the EU and world
imports recovered quickly. A particularly strong
import rise in China helped ease the recovery in East-
Asian countries.9 The rebound of world imports
starting in 2010 have boosted as well EU exports
(Figure 1.7).
Figure 1.7. World imports and EU exports from 2000
to 2012
Source: UN COMTRADE Note: Trade data for 2012 is still
incomplete at the time of drafting this report. Imports and exports
in current value. (EU imports excluded)
This question will be explored using an indicator for
competitiveness on world markets, the index of
revealed comparative advantage (RCA). The RCA
index compares the share of an EU sectors exports in
the EUs total manufacturing exports with the share
of the same sectors exports in the total
manufacturing exports of a group of reference
countries. RCA value which is higher than 1 means
that a given industry performs better than the
reference group and has comparative advantage,
European Commission (2012).
while value lower than unity indicates comparative
disadvantage.10
According to the RCA indices, some 15
manufacturing industries had comparative advantages
in 2009 and 2011.11 Two thirds of these are either
low-technology or medium-low-technology
industries. However, the EU has comparative
advantages in most medium-high-technology
industries as well as the high-technology sector of
pharmaceuticals (Figure 1.8)
Figure 1.8. EU comparative advantages in 2009 and
2011
The indicators above are calculated from trade data.
Even though the industries can be classified
according to technology intensities, it is hard to
measure the real sophistication of a country's
manufacturing using this kind of data, for at least two
reasons. The first reason has to do with the
difficulty of observing the quality or complexity of
the export products of a country or an industry.
Two products in the same sector or even two products
with the same customs code can have different degree
of complexity. Secondly, for any given good, trade
statistics do not provide information on the share
of value added produced domestically (i.e. the
domestic content of a country's exports). That
makes it difficult to tell for example if an industry in
a specific country is developing high-tech products or
merely assembles them. These limitations complicate
measurements and comparisons of industrial
competitiveness. They also require that the picture of
EU competitiveness based on RCA presented in
Figure 1.8 be extended to account for these two
additional indicators of international competitiveness.

Concerning the quality or complexity of a product,

a recent strand of literature interprets the
Source: UN COMTRADE competitiveness of an industry in a certain country
through its ability to produce relatively sophisticated
A comparison with major global competitors products. Two key concepts in this respect are the
(including BRIC) in sectors grouped according to diversification of the export mix of a given country or
technology intensities in Table 1.1, shows that EU, industry; and the sophistication or exclusiveness of
Japanese and US manufacturing industries have the export mix.13 Diversification by itself does not
RCAs in medium-high-tech sectors. Only Chinese indicate strong capabilities in an economy with
high-tech manufacturing has RCA of 1, 56.12 complex productive structures. It could very well be
that a country whose industries produce a large
Table 1.1 Revealed comparative advantages by number of products does so because the products are
technology intensities in manufacturing 2011 at the end of their life cycles, i.e. are standardised and
High Medium Medium Low can be produced at low costs. Countries with more
tech high tech low tech Tech complex productive structures will have industries
EU 0.85 1.14 0.89 1.01 that are able to produce more sophisticated and
Japan 0.73 1.59 0.86 0.16
US 0.88 1.22 0.96 0.68
Figure 1.9. Product complexity: comparison of
Brazil 0.32 0.76 0.87 2.50 advanced manufacturing goods with non-Key
China 1.56 0.72 0.85 1.29 Enabling Technologies
India 0.40 0.49 1.93 1.33
Russia 0.08 0.45 2.74 0.49
Source: UN COMTRADE

10
See Balassa (1965). A disadvantage with the measure is that it
can assume values between zero and infinity. See European
Commission (2010a) for an alternative specification that
Note: Product complexity is the average 2005-10. More density
constrains the index to range from -1 to +1 with positive
to the right means products in that category are more complex
values indicating revealed comparative advantages.
11 than average.
One should note that the manufacturing industries above are
represented for the two-digit level NACE classification. This is
a relatively high level of aggregation which includes a great
13
many industries. See Annex 1 for a description of the methodology for
12
These aggregates mask significantly differences not only calculating complexity of products. The description is based on
between the industries entering the aggregates but also Reinstaller, A. et .al (2012). See also Felipe et. al. (2012 pp.
between the different EU Member States. 36-68).

18
exclusive products.14 These countries have a
knowledge base or critical mass large enough to
produce sophisticated products.
The complexity of products can be illustrated by
comparing the type of products produced by
advanced manufacturing vis--vis the other
manufacturing sectors. Figure 1.9 shows how
complexity is distributed within the category
advanced manufacturing technologies and in those
sectors that do not belong to key enabling
technologies. In the first category most products are
more complex than average while the second shows a
more even distribution between complex and less
complex goods.15
Figure 1.10. Percentage of total products for which EU manufacturing industries have revealed comparative
advantages in different levels of complexity: averages 2005-2010 NACE Rev. 1
Source: Reinstaller et. al. (2012). BACI database. Note: Products are sorted according to their complexity score on the horizontal axes. The
average world market shares for each countrys products are shown on the vertical axes.
Calculating revealed comparative advantages for
products according to different degrees of
complexity, across EU manufacturing industries for
2005-2010 average value, shows how successful EU
manufacturing industries are in competing with goods
of different degrees of complexity.
The figure below focuses on EU products with
revealed comparative advantages (values of RCA
above 1). The results show that quite a high share of
products with RCA in the categories of basic and
fabricated metals (NACE 27 and NACE 28),
machinery and equipment n.e.c. (NACE 29), office
machinery and computers (NACE 30) and motor
vehicles (NACE 34) are products of higher than-
14
See Hausmann and Hidalgo (2011); Hausman, Hwang and
Rodrik (2007)
15
See also the discussion in European Commission (2013a).
19
average complexity. For example, more than 90% of
the products for which the manufacturing industry
medical, precision and optical instruments (NACE
33) have revealed comparative advantages are more
complex than the average products sold on world
markets by the same industries in other countries.
EU manufacturing industries have revealed
comparative advantages also for low complex
products such as tobacco (NACE 16), clothing
(NACE 18), leather (NACE (19) and wood (NACE
20) (Figure 1.10)
Given that competitiveness is a dynamic state, it is
interesting to see how the EU manufacturing
industries products change in terms of complexity
over time. EU manufacturers producing the most
complex items: chemicals (NACE 24), machinery and
equipment (NACE 29), medical, precision and optical
instruments (NACE 33) and motor vehicles (NACE
34) maintained their position in 2010 compared to
1995. Industries producing electrical machinery and
apparatus n.e.c. (NACE 31) and radio, TV and
communication equipment (NACE 32) managed to
upgrade their products, while industries producing
office machinery and computers (NACE 30) were not
able to upgrade their average complexity, (Figure
1.11.)16
16
The lines crossing zero at the horizontal and vertical axes
denote the average complexity of industries products in 1995
and 2010 respectively. A dot above the 45 degree line indicates
that an industry has managed to increase its average
complexity between 1995 and 2010.
Larger changes have taken place over time in
products supplied by BRIC countries. Products from
wood industries (NACE 20), radio, TV and
communication equipment (NACE 32), medical,
precision and optical instruments (NACE 33) and
furniture industries (NACE 36) have considerably
improved the average complexity of their products,
(Figure 1.12)17
Figure 1.11. Development of product complexity at EU
manufacturing industry level between 1995 and 2010
NACE Rev. 1.
Source: Reinstaller et. al. (2012). BACI database. Note: Products
are sorted according to their complexity scores 1995 and 2010
on the axes.
Even though industries in the BRIC countries
managed to upgrade their products considerably
between 1995 and 2010, the majority of industries in
these countries still produce less complex products
than their counterparts in the EU. In fact,
manufacturing industries in the EU have a high
degree of complexity. This is further confirmed by
the observation that the EU exported about 67% of
products with revealed comparative advantage in
2010. In comparison, the US only has a comparative
advantage in 43% of products, China in 54% and
Japan in 24%.18
The EU is a highly diversified economic area, which
is further confirmed by Figure A 2 to A6 in the annex
to this chapter. More industries in the EU than in
Japan and South Korea are able to secure big market
shares in a larger number of export products.
Manufacturing industries in the US are more
advanced competitors in this respect. China has
developed over time more industries able to produce
relatively complex products than before. Chinese
manufacturing industries are however still
predominantly competitive in product categories with
lower complexity.19 Reinstaller et al (2012) show that
17
The lines crossing zero at the horizontal and vertical axes
denote the average complexity of industries products in 1995
and 2010 respectively. A dot above the 45 degree line indicates
that an industry has managed to increase its average
complexity between 1995 and 2010.
18
See Reinstaller et al, 2012.
19
ibid
20
EU exporters, together with those in the US, Japan
and South Korea are more able to capture larger
shares of the world market by offering more
exclusive products which rely on a broader
knowledge base.20
The analysis of export complexity and exclusivity
feeds into the broader discussion of the upgrade of a
country's productive structure and comparative
advantages. It is discussed in the context of the
differences in structural change across countries in
chapter 2 of this report.
The issue of separating the domestic content of
production from foreign content is related to the
increased international fragmentation of production
which gives rise to increased intra-industry trade in
intermediate goods. In traditional trade statistics the
value of imported intermediate goods is included in
the export value of the final product that is exported.
One possibility is to adjust gross export flows for
imported intermediates by means of global input-
output statistics. The resulting exports only capture
Figure 1.12. Development of product complexity at
BRIC countries manufacturing industries levels
between 1995 and 2010
Source: Reinstaller et. al. (2012). BACI database. Note: Products
are sorted according to their complexity scores 1995 and 2010
on the axes.
the value added which is generated domestically in
the production of goods destined for export (see
Johnson and Noguera, 2012; Stehrer, 2012) but
exclude foreign value added associated with imported
intermediates. Value added export also exclude the
part of value added which is created domestically but
is used in domestic production.21
A related concept is value added content in trade.
This concept measures the domestic and foreign value
added embodied in a countrys gross exports. The
measure provides information on how much value
added from the exporting and other countries is
20
Regression analyses show that increases of product complexity
of EU manufacturing products are positively associated with
increases of world market shares, employment and value added
growth in the EU manufacturing sector. See Reinstaller et al,
(2012 pp 32-33).
21
See chapter 4 in this report.
embodied in a countrys gross exports. A large share 2007. An exception is Korea which hosts a large
of foreign value added content in a countrys exports number of Japanese multinational firms.23
is indicative of a less sophisticated part of the
Foreign value added embedded in gross exports can
Table 1.2. Domestic and foreign value added content of gross manufacturing exports by source country in 1995 and
2009 (%)
EU CHINA JAPAN KOREA US
1995 2009 1995 2009 1995 2009 1995 2009 1995 2009
Domestic 91.1 85.6 82.7 73.6 93.3 85.4 73.3 61.3 86.9 84.5
Foreign 8.9 14.4 17.3 26.4 6.7 14.6 26.7 38.7 13.1 15.5
EU 2.8 5.1 1.2 1.8 4.4 5.2 3.7 3.3
CHINA 0.3 2.3 0.4 2.4 1.7 6.7 0.4 2.5
JAPAN 1.0 0.7 3.8 3.3 6.3 4.7 2.2 0.9
KOREA 0.3 0.4 2.0 1.8 0.5 0.5 0.6 0.4
US 2.3 2.4 2.0 3.4 1.4 1.6 5.1 3.8
AUSTRALIA 0.2 0.2 0.5 1.3 0.3 0.9 1.1 1.8 0.1 0.2
BRAZIL 0.2 0.4 0.1 0.6 0.1 0.2 0.3 0.4 0.2 0.3
CANADA 0.4 0.4 0.4 0.5 0.3 0.3 0.7 0.5 1.8 2.0
INDONESIA 0.1 0.2 0.5 0.4 0.3 0.6 0.6 1.2 0.1 0.1
INDIA 0.1 0.3 0.1 0.3 0.1 0.1 0.2 0.3 0.1 0.3
MEXICO 0.1 0.2 0.0 0.2 0.1 0.1 0.1 0.2 0.7 1.2
RUSSIA 0.8 1.5 0.3 0.7 0.1 0.4 0.4 1.0 0.2 0.3
TURKEY 0.1 0.3 0.1 0.1 0.0 0.0 0.0 0.1 0.0 0.1
TAIWAN 0.2 0.2 1.8 1.8 0.3 0.4 0.6 0.9 0.5 0.3
Rest of world 2.8 5.0 2.9 7.1 1.7 5.2 5.4 11.9 2.4 3.8
Source: WIOD
production process, such as the assembly of a be broken down by source. This will show whether
product.22 Figure 1.13.compares foreign value added countries and their industries succeed in selling
of exports from China, the EU, Japan, Korea and the intermediate inputs to be used in the gross exports of
US from 1995 to 2009. It shows that foreign value other countries. The value added content of
added embedded in EU manufacturing exports is manufactured gross exports by source is closely
lower than that of the other global competitors. related to the measure of vertical specialisation
(Hummels et al, 2001). Between 1995 and 2009 when
Figure 1.13. Lower extent of foreign value added Chinese exports increased dramatically, EU value
embedded in EU manufacturing exports added in Chinese manufacturing exports increased
more than that of industries from other parts of the
world (Table 1.2). Japanese, Korean value added in
Chinese manufacturing exports decreased during the
same time. The increased presence of inputs from the
rest of the world in Chinese, Japanese and Korean
manufacturing gross exports suggest that there is a
strong inter-Asian production network masked in this
aggregate.
One reason for a relatively lower foreign content of
EU manufacturing gross exports is that most of the
value chains in which EU firms participate are
Source: WIOD. regional, i.e. within the EU. The manufacturing
aggregate masks differences across industries. The
The effects of the financial crisis on trade and global value chains involving EU industries producing
value chains are visible in the figure as the shares of chemicals, electrical equipment and transport
foreign value added content ceased to increase after equipment are more global, with a higher foreign
content of manufactured exports.

22 23
See Stehrer (2012) for an extensive discussion of these two OECD (2011). See also previous OECD Economic surveys for
concepts. Korea.

21
Another reason for a lower foreign content of exports
can be that the ability to produce most of the value
added content of high-tech production and exports
within a country can be an indication of complex
productive structures. A look at the domestic value
added content embedded in manufacturing high-tech
exports reveals that the EU, Japan and the US are
better able than Chinese or Korean counterparts to
source most of the input factors necessary for high-
tech production at home (Figure 1.14).
Figure 1.14. Lower extent of domestic value added
content in high-tech manufacturing exports from
China and Korea
Source: WIOD.
Figure 1.15. More favourable developments of unit labour costs in high- and medium-high-tech industries 2001-
2012
Source: Own calculations using Eurostat data. Note: Annual growth 2001-2012 in ULC based on employment (%). HT, HMT, LT and
LMT denote high-tech, medium-high-tech, low-tech and medium-low-tech manufacturing industries respectively.
The analyses in European Commission (2011b) show
that Chinese exports of high-tech manufacturing
depend to a large extent on high-tech intermediate
imports from other countries. This is evidenced by
relatively low values per unit of high-tech export
against relatively high values per unit of imported
intermediate goods.24
24
High-tech industries here include NACE 30 to 33, i.e. it
include also the medium-high-tech group NACE 31, as it is
22
1.3. DRIVERS OF SECTORAL COMPETITIVENESS
This section assesses sectoral performance looking at
drivers of external competitiveness. The development
over time of cost and price competitiveness is
analysed first. This is followed by analysis of
indicators of determinants of non-price
competitiveness.
Labour cost and productivity
Developments in labour costs should be assessed in
relation to labour productivity. A common measure is
unit labour cost (ULC), which is defined as the ratio
of labour compensation to labour productivity.
Labour compensation and labour productivity can be
measured either relative to the number of workers or
the number of hours worked. Increases in labour costs
exceeding labour productivity growth imply lower
profits on markets where the competition is intense
and where firms are price takers. Developments in
ULC can therefore be regarded as measures of cost
competitiveness on markets of non-differentiated
products. It should be noted that at given labour costs,
ULC developments are heavily influenced by
business cycle fluctuations impacting labour
productivity growth through larger variations in
production than in employment or hours worked.
For firms which produce homogenous goods and face
strong competition from low-cost countries, labour
costs are an important means to remain competitive.
ULC may however not be a good indicator for firms
which produce differentiated goods with some
characteristics that allow the firms some room of
based 2-digit NACE Rev. 1.1 classification and the
aggregation of industries in the World Input-Output Database
(WIOD).
manoeuvre to set the prices themselves. Such firms
producing goods with higher value added are more
frequently found in high-tech and medium-high-tech
manufacturing industries. Their goods are often
combined with some kind of services aiming at
satisfying demand for differentiated goods in high-
income segments of different markets. Labour
normally constitutes a smaller proportion of total
costs and input factors for such firms, rendering the
ULC less useful as a measure of competitiveness.
ULC based on the number of employees in
manufacturing and mining industries are compared
below. High-tech and medium-high tech industries
display lower ULC growth rates. ULC growth rates
for manufacturing are considerably lower than for
mining for the whole period and for different sub-
periods (Figure 1.15).
Comparisons of different industries in the EU provide
some insight into the competitiveness of these
sectors. It is however more meaningful to compare
developments of indicators of EU competitiveness
with the same indicators for industries from other
parts of the world which compete with EU firms.
ULC developments for aggregate EU and US
manufacturing are compared in Figure 1.16.25 Figure
1.19 shows that US unit labour costs advantages are
driven mainly by labour productivity. As can be seen
below, a larger fall in EU labour productivity growth
following the outbreak of the crisis is reflected in a
higher growth of unit labour costs (Figure 1.16).
Figure 1.16. Similar developments of ULC in EU and
US manufacturing industries
Source: Own calculations using Eurostat and Federal Reserve
data. Note: Growth rates in percent. ULC based on hours worked.
Most of the recent variations in EU manufacturing
unit labour costs are due to fluctuations in labour
productivity growth. Taking the analysis one step
further, faster decline of output relative to
employment during the slump accounts for most of
the losses in labour productivity in the EU at the start
25
It would be more interesting to compare ULC developments
between different types of EU and US manufacturing
industries. It is however hard to make these comparisons,
because of different industrial classifications.
23
of the crisis. Between the first quarters of 2008 and
2009, production decreased by 19% while hours
worked fell by 8% (Figure 1.17). This can partly be
explained by labour market rigidities in Europe on the
one hand and labour hoarding on the other (e.g.
enterprises avoid the higher cost of recruiting and
retraining when demand picks up by keeping skills
in-house in time of slump).
Figure 1.17. Fluctuations in EU ULC are mainly
caused by variations in labour productivity growth
Source: Own calculations using Eurostat data. Note: Growth
rates in percent. ULC based on hours worked.
Growth of labour productivity is important both for
price and non-price competitiveness. Labour
productivity, and especially multi-factor productivity,
is often seen as indicator of technical progress. An
increased labour productivity means more output is
produced with less labour, which can be due to
technological or organizational improvements and
other non-observable factors. Labour productivity
growth is often used as an indicator of price or cost
competitiveness as by increasing productivity firms
can lower their prices at given labour costs.
Between 2000 and 2011 labour productivity,
measured as value added per employee, grew faster in
high-tech manufacturing and knowledge-intensive
ICT sector. Some low-tech and medium low-tech
industries such as textiles and rubber and plastics also
performed relatively well and above the
manufacturing average. The lowest productivity
growth rates are observed in labour intensive services
(Figure 1.18)
The relationship between labour productivity growth
and market share gains is not straightforward. Firms
in industries facing tough competition from low-cost
producers (e.g. textiles and other low-tech sectors) are
forced to rationalize their production in order to
survive. Productivity growth in such a case may occur
together with a declining share of the world market.26
Therefore it is more informative to compare
26
In the worst case, firms are forced to close down non-
profitable plants and reduce the labour force in order to
rationalize. This would, if all other factors are equal, bring
about all things equal yield and an increase in d industry
productivity growth.

Figure 1.18. Highest labour productivity growth in
ICT manufacturing and pharmaceuticals
Source: Own calculations using Eurostat data. Note: Annual
growth in productivity per person employed 2000-2012 (%).
productivity growth rates with these of EU major
competitors.
Labour productivity growth in US manufacturing in
2000-2011 was 3.5% on average against 2.4% in EU
manufacturing.27 Large part of this difference has
occurred in the beginning of the millennium, even
though a larger decline of labour productivity in the
EU between 2008 and 2010 contributed to it too.
Figure 1.19 shows that during recessions
manufacturing employment (in hours worked) tend to
decline more in the US than in the EU. Therefore at
similar decline of demand and output manufacturing
labour productivity declines more in the EU.
27
Measured as changes from a quarter in one year relative to the
same quarter in the previous year.
24
The remainder of this section examines non-price
competitiveness of EU manufacturing. Accounting
for the determinants of non-price competitiveness
however, is a challenging task. There is a rather large
spectrum of factors which determine the good's
quality and its value for the customer.28
Figure 1.19. Manufacturing labour productivity in EU
and US
Source: Own calculations using Eurostat and OECD data. Note:
Annual growth in labour productivity per hours worked (%).
Innovation activities of firms resulting in product and
process innovations as well as marketing and
organization innovations are often regarded as non-
price competitiveness factors. The discussion in
European Commission (2010a) addresses also the
quality and variety of inputs such as intermediate
goods, service inputs, or the framework conditions
under which firms operate. The analyses here will
focus on human capital, physical capital, R&D and
innovation and the use of services inputs.29
Skills
Labour and skills are not perfectly mobile, i.e. they
cannot be moved across sectors without cost. The
labour force consists of individuals possessing
different types of skills and levels of education. This
heterogeneity makes hiring and firing costly as they
entail search and transactions costs. Highly educated
labour with a certain set of skills can be difficult to
find within any given period of time. This makes
firms reluctant to make this kind of labour redundant
during recessions. Adding to this reluctance are the
sometimes firm specific skills that the labour force
acquires within the firm.
These characteristics of the labour force mean that it
is necessary to discuss skills and human capital as
an input factor which can explain differences in
growth between countries. Human capital is not
easily measured. An often used proxy for
28
See the discussion of how to define and measure for example
the quality of goods in NUTEK (1997).
29
Some of these, and other, non-price competitiveness factors
are analysed by means of regression analysis in European
Commission (2010a).
accumulated knowledge is educational attainment. It
is an imperfect measure since it is not capable of
taking into account the whole stock of knowledge
built up by skills and experience acquired after school
from vocational or on-the-job training and learning by
doing.30 This indicator has however the advantage of
Figure 1.20. Skill and knowledge intensities (% of total
employment)
Source: Own calculations using Eurostat's labour force survey
data.
being easily available. It is used here to analyse the
distribution of employment by education across
30
For a discussion of proxies for human capital in empirical
studies, see Greiner, Semmler and Gong (2005). On different
ways of measuring the stock of human capital, including a
discussion on the limitations of educational attainment as a
proxy for human capital, see OECD (1998).
25
sectors based on the International Standard
Classification of Education (ISCED)31.
The market and non-market services sectors of
education, information and communication and
financial activities are the most human-capital
intensive. Manufacturing industries which produce
goods that require a relatively high share of high-
skilled labour are pharmaceuticals, refined petroleum
products and computer, electronic and optical
industries. Around 50% of the labour force in
pharmaceutical firms has tertiary education. The
smallest share of low-skilled workers is found in
financial services where only 5% of the labour force
have no educational qualifications beyond primary
level. More than 25% of the workforce in
manufacturing industries producing chemicals, other
transport and tobacco are also classified as highly-
skilled.
Low-technology manufacturing sectors like textiles,
clothing and leather have small shares of highly
skilled labour, as do labour-intensive service sectors
such as hotels and restaurants and agriculture and
forestry (Figure 1.20)
Investment
Investment in physical capital increases output
capacity of firms and their labour productivity. It also
improves total factor productivity by bringing
technology, innovation and intangibles, thereby
facilitating reorganisation and adaptation of the
production process to shifts in consumer demand.
Conversely, lower investments today impacts
negatively not only current growth performance, but
also future growth prospects through a lower capital
stock but also through a lower future innovation and
productivity growth, as much of the R&D and
innovation is embedded in physical capital.32
The investment ratios presented in Table 1.3. below
are defined as the ratio of gross fixed capital
formation (GFCG) to value added.33 Sectors with a
high share of large capital intensive firms such as
transport equipment, electricity and gas, water supply,
transportation and storage as well as real estate
activities have high investment ratios. The statistical
classification with aggregation of sectors sometimes
distorts the picture. The tobacco industry, which is
capital intensive and dominated by large firms, is
grouped together with food and beverages industries.
31
ISCED identifies levels of education from 0 to 6 and is used to
measure the proportions of low-skilled, medium-skilled and
high-skilled labour for each sector (see annex 3 for definitions)
32
See for example the discussion in European Commission
(2010a).
33
Fixed assets; buildings, machinery and equipment, transport
equipment, office machinery and hardware, software and
intangible fixed asset are included in this aggregate.
That lowers the average investment ratio for this
aggregation of sectors. The investment ratios are
relatively stable over time with some significant
exceptions. Following the financial crisis, production
in petroleum industries fell by some 40% in 2009
while investments remained as high as the previous
Table 1.3. Investment ratios in 20 Member States
year. A possible reason for maintaining a high
investment against drop in demand could be that the
downturn was perceived as temporary, that the prices
of investment goods declined, or the investment cycle
in extracting industries has longer-term time horizon
and is less responsive to short-term fluctuations of
demand, or for other strategic reasons. In any case,

2007 2008 2009 2010 2011

TOTAL Total 0.25 0.24 0.22 0.22 0.22
A Agriculture, forestry and fishing 0.37 0.39 0.39 0.34 0.35
B Mining and quarrying 0.25 0.22 0.28 0.24 0.27
C10-C12 Food, drinks and tobacco products 0.19 0.20 0.17 0.17 0.21
C13-C15 Textiles, clothing, leather and footwear 0.12 0.12 0.10 0.11 0.15
C16-C18 Wood, pulp and paper and printing 0.20 0.21 0.17 0.17 0.21
C19 Refined petroleum products 0.35 0.36 0.61 0.41 0.27
C20 Chemicals 0.19 0.21 0.19 0.16 0.26
C21 Pharmaceuticals 0.15 0.15 0.14 0.13 0.18
C22_C23 Rubber and plastics 0.20 0.20 0.16 0.16 0.21
C24_C25 Basic metals and metal products 0.18 0.19 0.18 0.17 0.21
C26 Computers, electronic and optical products 0.18 0.18 0.18 0.19 0.28
C27 Electrical equipment 0.13 0.13 0.12 0.11 0.14
C28 Machinery and equipment n.e.c. 0.12 0.13 0.12 0.10 0.13
C29_C30 Transport equipment 0.20 0.23 0.23 0.18 0.32
Furniture, repair and installation of machinery and
C31-C33
equipment 0.11 0.11 0.10 0.11 0.13
D Electricity and gas 0.37 0.36 0.38 0.39 0.37
E Water supply 0.49 0.48 0.45 0.46 0.43
F Construction 0.16 0.15 0.10 0.11 0.10
G Wholesale and retail trade 0.12 0.12 0.10 0.10 0.10
H Transportation and storage 0.37 0.39 0.36 0.37 0.37
I Accommodation and food service activities 0.12 0.12 0.10 0.10 0.12
J58-J60 Publishing, motion picture and broadcasting 0.23 0.23 0.21 0.22 0.26
J61 Telecommunications 0.29 0.29 0.25 0.27 0.28
J62-J63 Computer programming and consultancy activities 0.16 0.17 0.15 0.16 0.18
K Financial and insurance activities 0.11 0.14 0.12 0.11 0.10
L Real estate activities 0.73 0.68 0.61 0.60 0.58
Legal and accounting activities and architectural
M69-M71
and engineering activities 0.09 0.09 0.08 0.08 0.10
M72 Scientific research and development 0.27 0.26 0.26 0.28 0.29
Advertising and market research, other professional
M73-M75
services, scientific and veterinary activities 0.09 0.09 0.08 0.09 0.12
N Administrative and support service activities 0.35 0.34 0.26 0.28 0.31
O Public administration and defence 0.27 0.27 0.27 0.25 0.21
P Education 0.09 0.09 0.09 0.09 0.09
Q86 Health care 0.12 0.12 0.11 0.11 0.10
Q87-Q88 Residential care activities and social work activities 0.13 0.12 0.11 0.11 0.07
R Arts, entertainment and recreation 0.35 0.34 0.33 0.31 0.21
S Other service activities 0.09 0.09 0.08 0.08 0.07
Source: Own calculations using Eurostat data.

26
the investment ratio in this industry increased to 0.61
in 2009. It should be noted that data on gross fixed
capital formation are not complete. Table 1.3 is based
on data for only 20 Member States,34 for 8 of which
2011 data is missing.35
economic activity. Public expenditures in terms of
sectoral R&D are not reflected in the data.
Figure 1.21. US firms spend more on R&D in almost
all sectors

R&D and innovation

R&D and innovation are indicators of non-price
competitiveness describing attempts by producers to
increase their competitiveness by improving supply
side conditions as well as trying to influence demand
for their products. R&D expenditures can improve the
supply by introducing technology which improves the
production process and lower production costs.
Outcomes of R&D can also be innovations in the
form of new, improved or differentiated products
which can increase the competitiveness of firms by
making demand for their products less price elastic.36
The adoption and use of technology determines how
efficiently input factors are combined in order to
achieve growth in the long run. The indicators below
describe the technology in the EU industries from
different angles. The indicators represent different
stages of the R&D&I process. R&D expenditures can
be regarded as input indicators while patents and
firms introducing new and/or improved products to a
higher extent measure outputs of the R&D&I
processes.
Due to insufficient coverage of R&D statistics across Note: The EU is represented by 17 countries: Austria, Belgium,
sectors and countries since 2007, the discussion in Czech Republic, Denmark, Finland, France, Germany, Greece,
this section cannot include more recent period. This Hungary, Ireland, Italy, Netherlands, Poland, Portugal, Spain,
may not be as serious a drawback as it seems. The Sweden and the UK. The industries are classified according to
ISIC Rev 3.1.
process from investing in R&D to developing new
products can be very long, especially in industries
such as pharmaceuticals where the process also This aggregate for EU manufacturing sectors is
includes rigorous and carefully regulated testing of compared with US manufacturing (Figure 1.21). The
the products. It may therefore be that the data comparison shows that R&D intensity in US
represented in the figures below more accurately manufacturing is higher, not due to differences in
describes the situation today than more recent data industrial structures but to an overall smaller
would do. Data on R&D today should be seen as investment in R&D in the EU across all sectors.
indication of future technology and innovation Patent statistics are often used to compare countries
results. and industries knowledge output. Even if indicators
EU R&D expenditures represented 1.85% of GDP in of patenting and the underlying statistics are subject
2007 against 2.7% in the US. The bulk of the to uncertainty and even bias, the information is of
difference between the EU and the US is found in interest.37
private enterprise R&D. An EU aggregate has been
formed in order to analyse R&D intensities by sectors
(i.e. R&D expenditures relative to value added). The
aggregate represents more than 80% of total R&D
expenditures in the EU. The analysis focuses on
Business enterprise R&D expenditures (BERD) by

34
BE, CZ, DK, DE, EL, ES, FR, IT, CY, LT, LU, HU, NL, AT,
PL, PT, SI, SK, FI and SE.
35
DE, ES, CY, NL, AT, PL, PT and SI.
36 37
See European Commission (2010a), p 123, and related Griliches (1990) discusses a number of issues related to
references, for a discussion of firms' attempts to differentiate patents, including the advantages and drawbacks. See also
products in order to increase their competitiveness Pavitt (1985), Silverman (2002) and Griliches (1984).

27
Figure 1.22. Relatively lower patenting by EU high-tech industries

Source: Own calculations using Eurostat data. Note: The aggregate World includes Iceland, Lichtenstein, Norway, Switzerland, Turkey,
Russia, South Africa, Canada, the US, Mexico, Brazil, China, Japan, South Korea, India, Israel, Taiwan, Singapore, Australia and New
Zealand.

Values larger than 1 indicate that the EU industry has

Patent statistics reflect the output of the research
a patent specialisation relative to the world. The
process undertaken by firms. The statistics provide
indicator shows that EU manufacturing industries
information on a large number of manufacturing
perform better than the world in a number of
sectors and the coverage over time allows trends and
industries. However, many high and medium-high-
correlations with other economic developments to be
technology industries such as pharmaceuticals, office
analysed. As data are available for many countries, it
machinery electrical equipment industries perform
is possible to calculate the performance of an EU
relatively worse than the world (Figure 1.22)38.
sector relative to say, global performance. The
measure is calculated in the same way as RCA
indices for manufacturing exports (see the annex for a 38
It should be noted that the indicator is based on patent
complete definition of the measure). applications to the EPO. The indicator might therefore be
biased in favour of EU manufacturing industries as there is a

28
Other things equal, a lower-than-average patenting of
EU manufacturing firms implies that EU industries
are less able to develop new and/or improved
products or production processes. This could translate
into future losses of competitiveness.
Firms engage in product innovation in order to
develop new or improve existing products. The
Figure 1.23. More innovative enterprises in manu-
facturing industries than in mining and service
industries
Source: Own calculations using Eurostat data. Innovative
enterprises as a percentage of total enterprises in the 2010 CIS
innovation survey.
purpose is to produce products that have certain
qualities that differentiate them from their
competitors. If they succeed they will face less elastic
demand thus having some ability to set their own
tendency for non-EU industries to patent relatively less
frequently at the EPO than at the USPTO. Triad patent families
for industries which could take this bias into account were not
available at the time of drafting of this report.
29
prices and be less reliable on labour costs and input
prices to compete. By engaging in process innovation
firm aim at implementing new production processes
that increase their productivity and/or lower their
production costs. Firms also engage in organisational
and marketing innovations to the same end.
EU manufacturing industries are more prone to
engage in innovation activities than services
industries.39 This is confirmed by the number of
innovative enterprises by sector as well as by the
number of innovations according to data from
Community Innovation Survey (CIS). Firms
producing pharmaceuticals, tobacco, computers,
chemicals and beverages have relatively higher shares
of innovative enterprises in all enterprises in the 2010
CIS. Pharmaceutical, ICT and chemical firms were
most successful in bringing new or improved
products to market between 2008 and 2010 according
to the Community Innovation Survey. It should
however be noted that innovation comes easier for
firms in some industries than others. Improving a
beverage or tobacco product by introducing a new
flavour is probably easier and less costly than
developing a new car model.
Few firms in low-tech manufacturing industries such
as clothing, wood and leather, construction industries
and in service industries (administration, hotels and
restaurants) are engaging in innovative activities
(Figure 1.23)
Services industries engage in process innovation to a
higher extent than product innovation. Market
services engage in process innovation almost to the
same extent as manufacturing firms (Figure 1.24)
The use of services in manufacturing
Manufacturing firms have increasingly used services
over time. This shows up on the input side as well as
the output side. They have increased their services
intensity in order to increase their productivity and
thereby also their competitiveness. Manufacturing
firms use services to differentiate their products from
their competitors.
On the input side, manufacturing firms use of
intermediate services has increased over time. The
increase has been most pronounced for services
provided by knowledge-intensive business services
firms (European Commission 2010). But
manufacturing firms are also producing more of the
services in-house. This turns up in the increased share
of employees with services-related occupations over
39
The figures are calculated as averages for different sectors
across the EU countries. The interpretation of the figure should
be treated with caution since there are gaps in the dataset. The
averages for tobacco, administration, accommodation and food
and real estate activities are based on ten or fewer
observations.
time. Having access to this kind of labour makes it and refined petroleum where 25% of services value-
easier for manufacturing firms to provide their added is imported. The largest total share of services
Figure 1.24. Pharmaceutical and ICT firms more successful in innovation

Source: Own calculations using Eurostat data. Data from the 2010 Community Innovation Survey.

physical goods with services characteristics and to Figure 1.25. Manufacturing firms services innovations
engage in services innovations. High-tech in 2010
manufacturing firms producing ICT and electronic
and optical equipment and pharmaceuticals are the
most frequent innovators in EU manufacturing. But
firms in the refined petroleum and coke sectors are
also responsible for a relatively high level of services
innovations.40 Low-technology industries engage in
services innovation much less than the EU
manufacturing average (Figure 1.25)
On the output side, services have increased over time
as a share of manufacturing output (European
Commission 2011). But manufacturing firms not only
produce more pure services than previously. The
contents of manufacturing goods have changed as
more and more services are embedded in physical
Source: Own calculations using Eurostat data. Note:
products. Manufacturing enterprises that developed services innovations as
percentages of total enterprises in the CIS innovation survey 2010.
The latter trend is a natural consequence of Data are not available for Denmark, Germany, Greece and the
manufacturing firms trying to differentiate their UK.
products. This is a response not only to intensified
competition from low-cost producers in emerging value added embedded in exports is also to be found
countries but also attempts to satisfy increased in industries producing coke and refined petroleum
demand for varieties of goods which increase as (Figure 1.26).
incomes rise. Upgrading the products may also make One of the most prominent characteristics of
customers willing to pay a premium for them if the increased globalisation is the way production
products are perceived to be of high enough quality. processes are sliced up between different locations,
This makes demand for these products less price according to their comparative advantages.
elastic. EU manufactured exports consist to an Technological progress, especially in
increasing extent of embedded services. Domestic communications, and lower transportation costs are
services account for most of the services value-added major factors behind the emergence of global value
around 90% across all industries, except in coke chains (GVCs). Firms participate in GVCs in order
to increase their competitiveness or better satisfy
demand in different foreign markets. Focusing on the
40
See also Lofalk (2013) for an analysis of the servicification of first of these motives, engagement in GVCs may
Swedish manufacturing. The industries producing refined increase a firm's competitiveness by enhancing access
petroleum and coke are among the most servicified
manufacturing industries. to cheaper or higher-quality intermediate inputs
(OECD 2013). Industries in OECD countries using a

30
higher share of imported intermediates display on
average a higher productivity. The effects arise
mainly in three ways. Firstly through lowering prices,
as more intermediate imports lead to stronger
competition among intermediate producers. Secondly
through increasing the supply of varieties of
intermediates as imports grow. Finally through
increased productivity as new imported intermediates
may be more suited for the technology of final goods
destined for the foreign markets. Gaining access to
foreign knowledge by using imported intermediates
may also lead to higher innovation as the firms
knowledge bases increase (OECD 2013).
CONCLUSIONS AND POLICY IMPLICATIONS
This chapter uses a number of advanced indicators of
international competitiveness to provide insights into
the strengths and weaknesses of EU manufacturing
and to draw implications for EU industrial policy. It
finds that the EU has comparative advantages in most
of its manufacturing sectors. These sectors include
such vital high and medium-high tech sectors as
pharmaceuticals, chemicals, vehicles, machinery,
other transport equipment (which includes airspace),
but also low and medium low tech sectors such as
food, beverage, tobacco, paper and plastic. On the
downside, in the high-tech sectors Europe has
comparative advantage only in pharmaceuticals while
EU electrical equipment and computer, electronic and
optical products lag behind in international
competitiveness.
The EU has comparative advantage in the broad
category of medium-high tech industries (1.14), but it
is smaller than that of Japan (1.59) and the US (1.22).
The development of RCA overtime shows that China
Figure 1.26. Services value added in EU
manufacturing gross exports 2009
Source: WIOD.
is quickly gaining ground. Actually only China has
comparative advantage in the aggregate broad
category of high-tech industries (1.56) leaving far
behind in high-tech export specialization the US
(0.88), the EU (0.85) and Japan (0.73).
This is the state of play of competition in broad
categories of industries grouped according to
31
technology intensities. Taking the analysis further
down to product level, however, presents Europe into
much better position on the world market. In 2010
67% of European exports have revealed comparative
advantage, while China has comparative advantage in
54% of products, and US and Japan export
respectively 43% and 24% of their products with
comparative advantage.
Indicators of revealed comparative advantage (RCA)
are derived from trade data. They do not provide
information about the sophistication of the export
product and how much of it is produced in the export
country (i.e. the domestic share of export value
added). Looking at the complexity of EU exports
shows that the sectoral comparative advantages of EU
industries even in the medium-high tech group is
based on higher complexity (knowledge intensity) of
exports than the average. For instance in medical
precision and optical instruments 90% of products
with RCA are more complex than the average
products exported by other countries. The analysis
shows that the present comparative advantages of the
EU industries are a result of maintaining and
upgrading the sophistication of EU exports over the
last 15 years. At the same time the emerging
industrial powers (e.g. BRIC) have achieved much
faster upgrade of their exports, but are still lagging
behind the EU industries in sophistication of exports.
Even though China has RCA in the high-tech
category, it is still based on products with lower
complexity.
The analysis of exports in value added provides new
insights to the international performance of EU
manufacturing. It shows that advanced economies
have higher domestic content of exports as thanks to
their strong industrial base they can afford to supply
domestically at competitive terms most of the inputs
needed for their exports. Thus the domestic value of
exports of EU, US and Japan is around 85 %, while
that of China is 73.6% and that of Korea is 61.3%.
Chinese high-tech exports for instance seem to rely
heavily on high-tech imports of intermediates. For the
sector of electrical equipment for example China's
market share in gross exports is 9 p.p. higher than its
market share in exports in value added.41 This is
evidenced by the relatively low values per unit of
finished high-tech exports against the relatively high
unit values of imported high-tech inputs.
About 20% of the foreign inputs in Chinese exports
comes from the EU, which is higher than US and
Japan (about 13% each). This is a result of faster
growth of EU share during the last 15 years of
Chinese export boom.
These indicators present the EU position in the global
competition in manufacturing goods. They however
cannot inform policy about why EU is there and
41
See Chapter 4 for details
whether and how it can perform better. Therefore the
analysis looks at those factors that can explain
industrial performance and if properly targeted by
policy can improve it. Competitiveness is above all a
result of productivity gains. Accordingly, the analysis
departs from labour cost and productivity of EU
manufacturing, but goes further to explore such
determinants of total factor productivity as input of
skills, R&D and innovation intensities, fixed capital
formation and the contribution of services to
manufacturing competitiveness.
It shows that between 2000 and 2011 labour
productivity in the US has grown by 3.5% in average
against 2.4% in the EU. This is largely explained by
the fact that during downturns the US seems to do
better in labour productivity growth. One reason is
that employment in the US adjusts faster than
employment in EU to shrinking demand for goods
and services. Productivity is the major driver of the
US superior performance vis--vis Europe in terms of
unit labour cost (ULC), which is one of the common
explanatory indicators of cost and price
competitiveness. A possible implication of this
comparison is that the EU labour market needs to
gain flexibility in order to allow faster and more
efficient adjustment of labour to shifts in demand.
Labour market rigidities are often explained by
employment protection. During the slump however
employment adjustment lagged behind drop in
demand not just because employers could not lay-off
workers, but because in the technology-intensive
sectors they chose to keep them to avoid the cost of
re-hiring and re-training when demand picks up.
Therefore shortage of skills and resultant hoarding of
labour may be additional reason explaining why
labour demand response to decline in output is more
sluggish in Europe than in the US. Chapter three
looks into more depth at the efficiency and
productivity deficits of EU manufacturing and the
relevant policy responses.
Labour costs however have decreasing weight in EU
manufacturing competitiveness for two reasons. First,
the analysis shows that EU exports rely mainly on
knowledge-intensive products rather than on low-tech
labour intensive products. On the other hand
emerging economies are catching up fast not only in
the level of technology but also in terms of wages.
Therefore what is more important for Europes
competitiveness in the present day global supply
chains is total factor productivity (TFP). It accounts
32
for that part of GDP growth which cannot be
attributed to measurable factor inputs, and is
explained by skills, technology and process
innovation, and investment in intangibles.
US private spending on R&D (as a share of GDP) is
almost 1.5 times that of the EU (2.7% in the US vs.
1.85% in the EU). A sector break-down indicates that
this is not a result of differences in industrial
structures or US specialization in knowledge-
intensive sectors, but to an overall underperformance
of EU sectors in terms of R&D investment across all
sectors.42 The output of research is new products,
new technologies, new materials and processes. A
rough indicator of this output is patents. The chapter
documents that in a number of high and medium-high
technology industries (such as pharmaceuticals,
optical equipment, electrical equipment, medical and
surgical equipment, telecom and office equipment,
radio and TV and accumulators and batteries), the EU
is lagging behind in patenting. As the RCA indicators
show, EU export performance depends crucially on
some of these sectors. It may be hard to preserve EU
current comparative advantage in these industries if it
loses its technology lead as indicated by patent data.
Another problem is that the way of EU research from
the lab to the market seems to be more difficult than
that of major competitors. This is an important
problem which deserves a more detailed study. This
report is trying to look for explanation in Chapter 4
and Chapter 5.
The implications of EU exports complexity for
industrial policy is that targeting only high-tech
sectors might be less rewarding than increasing the
share of knowledge intensive products in all tradable
sectors including medium-low and medium-high tech
sectors. Moreover some of the labour intensive
sectors with lower knowledge intensities may be
better suited to tackle EU's unemployment challenges
than the high-tech sectors. About 40% of EU
manufacturing employment is in low-tech sectors.
Therefore the policy priority attached to key-enabling
technologies which leads to new materials and
products in all manufacturing sectors has strong
potential to upgrade EU competitiveness not only in
the high-tech sectors but also in the traditional
industries. Chapter 5 in this report analyzes EU
performance and prospects in the competition in
KET-based products.
42
See Chapter 4 of this report for further details.
 REFERENCES
Balassa, B., Trade Liberalisation and Revealed Comparative Advantage, Manchester School of Economic and
Social Studies, 1965, Vol. 33, pp. 99123.
European Commission (2009), EU Industrial Structure 2009, Performance and competitiveness, Enterprise and
Industry DG, Office for Official Publications of the European Communities, Luxembourg.
European Commission (2010a). Product Market Review 2010-2011. European Economy 8. Directorate-General
Economic and Financial Affairs, European Commission.
European Commission (2010b). European Competitiveness Report 2010
European Commission (2011a). European Competitiveness Report 2011
European Commission (2011b), EU Industrial Structure 2011, Trends and Performance, Enterprise and Industry
DG, Office for official publications of the European Communities, Luxembourg.
European Commission (2012). European Competitiveness Report 2012. Reaping the benefits of globalisation
European Commission (2013a). Short-term Industrial Outlook, January 2013. Directorate-General Enterprise and
Industry
European Commission (2013b). Short-term Industrial Outlook, July 2013. Directorate-General Enterprise and
Industry
European Commission (2013c), EU Industrial Structure 2013, Trends and Performance?, Enterprise and Industry
DG, Office for official publications of the European Communities, Luxembourg.
Felipe, J. et. al. (2012). "Product complexity and economic development", Structural Change and Economic
Dynamics, 23 (2012), 36-68.
Foray, D., David, P. A. & Hall, B."Smart Specialisation - The Concept", Knowledge Economists Policy Brief
no. 9, June 2009, DG RTD.
Greiner, A., Semmler, W. & Gong, G. (2005). "The forces of economic growth - a time series perspective".
Princeton University Press, Section 4.3.
Griliches, Z. (ed.) (1984), "R&D, Patents and Productivity" The University of Chicago Press, International
Monetary Fund, Balance of Payments and International Investment Position Manual, sixth edition.
Griliches, Z. (1990), Patent statistics as economic indicators: a survey, Journal of Economic Literature, Vol.
XXVIII, pp. 16611707.
Hummels, D., Ishii, J., and Yi, K.M., (2001), 'The Nature and Growth of Vertical Specialization in World Trade',
Journal of International Economics, Vol. 54 No. 1.
OECD (1998), Human Capital Investment An international comparison, Center for Educational Research and
Innovation, Paris.
OECD (2011). OECD Economics Surveys: Korea 2010. OECD, Paris.
Pavitt, K. (1985), Patent statistics as indicators of innovative activities: possibilities and problems,
Scientometrics, Vol. 7, pp. 7799.
Silverman, B. (2002), Technological resources and the logic of corporate diversification, Routledge, Chapter 4.
Reinstaller, A., Hlzl, W., Kutsam, J., and Schmid, C., (2012), The development of productive structures of EU
Member States and their international competitiveness, Report prepared under Specific Contract No SI2-
614099 implementing the Framework Contract ENTR/2009/033 for the European Commission, DG
Enterprise and Industry.
Hausmann, R., Hidalgo, C. A. (2012), The network structure of economic output, Journal of Economic
Growth, 16, pp. 309-342.
Hausmann, R., Hwang, J., Rodrik, D. (2007) What you export matters, Journal of Economic Growth, 12(1),
pp. 1-25.
NUTEK (Swedish National Board for Industrial and Technical Development) (1997). "Market Shares, Relative
Prices and Quality. A study of Sweden's international competitiveness".NUTEK B 1997:9, Stockholm.
Stehrer, R. (2012). Trade in Value Added and the Value Added in Trade. WIOD working Paper 8 (2012) 1-19.

33
 ANNEX 1
DIVERSIFICATION AND UBIQUITY OF PRODUCTS
The analytical approach is based on Hidalgo et al. (2007) and Hidalgo and Hausmann (2009). This approach uses
trade data to construct measures for the diversification of an economy and the sophistication of the products it
exports. Data from the Base pour lAnalyse du Commerce International (BACI) database developed at the
Centre dtudes Prospectives et dInformations Internationales (CEPII) have been used. The dataset contains
data for 232 countries and 5,109 product categories classified using the Harmonized System at the 6-digit level.
The data cover the years 1995 till 2010. The methodological approach aims to capture, in a few indicators, the
productive capabilities in an economy.43 Figure A 1shows how these indicators are constructed. A country is
linked to a product if the countries have revealed comparative advantage (RCA) in this product. The implicit
assumption is therefore that a country disposes of capabilities or factor endowments that convey a competitive
advantage in this product.

Figure A 1. Diversification and ubiquity of countries and products

If the matrix shown in the figure above is summed up column-wise over products p one obtains a measure for
the diversification of a country c.

diversification

Where k is the measure of diversification. M is an indicator which assumes the value of one (1) if RCA >1 for a
country c exporting a product p.
If on the other hand the matrix is summed up row-wise one obtains a measure for the ubiquity of comparative
advantage in the trade of a specific product p. This measure tells us how many countries c have a comparative
advantage in trading this product.

ubiquity

By combining these two indicators it is possible to calculate through recursive substitution how common
products are that are exported by a specific country,

for

and how diversified the countries are that produce a specific product

for

43
A short an intuitive description of the methodology is available in European Commission (2013a)
http://ec.europa.eu/enterprise/policies/industrial-competitiveness/competitiveness-analysis/index_en.htm.

34
n in (3) and (4) denotes the number of iterations in the computations. See also Table A 1. If formula (3) goes
through an additional iteration the indicator now tells us how diversified countries are that export similar
products as those exported by country c. An additional iteration for formula (4) then tells us how ubiquitous
products are that are exported by product ps exporters. Table A 1 gives an overview on how the indicators can
be interpreted.44 Only the first three iterations of the indicator are presented below. The indicators k_(p,max)
and k_(c,max) provide for any product p, (k_(p,max)), its level of complexity, and for any country c, k_(c,max),
the level of complexity of the productive structures of its economy.
These two indicators are calculated by going through as many iterations necessary until the ranking of the
countries and the products in terms of the k_(p,max) and k_(c,max) values do not change anymore. The number
of iterations necessary to obtain this convergence may thus vary from year to year.

Table A 1. Interpretation of the indicators calculated using the Method of Reflections, first three pairs

n country
0 : number of products exported by country c,
diversification
How many products are exported by country c?
1 : average ubiquity of products exported by
country c
How common are the products exported by
country c?
2 : Average diversification of countries with a
similar export basket as country c How
diversified are countries exporting similar
products as those exported by country c?
Source: Abdon et al. (2010), p. 8, following Hidalgo - Hausmann (2009), Supplementary material p.8
product
: number of counries exporting product p,
ubiquity
How many countries export product p?
: Average diversification of the countries
exporting product p
How diversified are the countries exporting
product p?
: Average ubiquity of the products exported by
countries exporting product p
How ubiquitous are the products exported by
product ps exporters?

The assumption underlying this analytical framework is that countries need a large set of complementary and
non-tradable inputs. Hausmann and Hidalgo refer to this as capabilities (see also Hausmann and Hidalgo 2011).
If countries differ in these capabilities and products differ in the type of capabilities that are needed to produce
and successfully trade them, countries with more capabilities will be more diversified. On the other hand,
products that require more capabilities will be successfully exported only by those countries that have these
capabilities, and as a consequence they will be less ubiquitous.
The indicators therefore capture on the one hand the variety of goods produced by an economy and to what
extent this product mix represents a unique source of comparative advantage for the economy. They do so by
conceiving these relationships as a network and by expressing the properties of each node in the network as a
combination of the properties of all its neighbours. This approach therefore exploits information from the global
trade network to construct indicators that capture important aspects of the level of economic development and
the competitiveness of economies by exploiting the fact that the economic fortunes of countries are intertwined
via trade, foreign direct investment, and financial capital flows.
The supply of products in one country is highly dependent on economic activities in multiple foreign countries
and changes in production networks spread across countries and continents. When countries and regions
transform as a result of economic, technological, political, or institutional change, the nature of foreign trade
changes as well, and trade data therefore capture such changes.

44
Higher iterations than those presented in the table are increasingly difficult to interpret.

35
 ANNEX 2
WORLD EXPORT SHARES AT PRODUCT LEVELS OVER
PRODUCT COMPLEXITY BY NACE, EU27 AND
COMPETING COUNTRIES
Figure A 2. World export shares at the product level over product complexity by NACE sector, EU 27

Source: Reinstaller et. al. (2012). BACI database

Figure A 3. World export shares at the product level over product complexity by NACE sector, US

Source: Reinstaller et. al. (2012). BACI database

36
Figure A 4. World export shares at the product level over product complexity by NACE sector, Japan

Source: Reinstaller et. al. (2012). BACI database

Figure A 5. World export shares at the product level over product complexity by NACE sector, Korea

Source: Reinstaller et. al. (2012). BACI database

37
Figure A 6. World export shares at the product level over product complexity by NACE sector, China

Source: Reinstaller et. al. (2012). BACI database

38
 ANNEX 3
DEFINITIONS OF MEASURES AND CLASSIFICATIONS
USED
The patenting, i.e. the number of patent applications, of a given EU manufacturing industry relative to total EU
manufacturing patent applications are compared to the number of patent applications of the same industry in the
world relative to the number of total patent applications in manufacturing in the world.45 The indicator, PAT,
measures the EU manufacturing industries relative patenting performance:

where:
PATiEU: number of patents filed by EU industry i
i PATiEU: number of patents filed by all EU

manufacturing industries

PATiWorld: number of patents filed by World industry

i
i PATiWorld: number of patents filed by all
manufacturing industries in the World

Box A.1: Using International Standard Classification of Education to define skill

categories
The International Standard Classification of Education (ISCED) differentiates seven levels of education.
Level 0: pre-primary
Level 1: primary education
Level 2: lower secondary
Level 3: upper secondary
Level 4: post-secondary non-tertiary
Level 5: first stage of tertiary education
Level 6: second stage of tertiary education.
The publication has aggregated the levels in three categories so that total employment in each sector can be
broken down in three skill categories instead of seven:
Low skilled: Level 0, Level 1 and level 2
Medium skilled: Level 3 and level 4
High-skilled: Level 5 and level 6

Manufacturing industries classified according to technological intensity (NACE Revision 2)

High-technology manufacturing
21 Manufacture of basic pharmaceutical products and pharmaceutical preparations
26 Manufacture of computer, electronic and optical products
Medium high-technology manufacturing
20 Manufacture of chemicals and chemical products
27 to 30 Manufacture of electrical equipment, Manufacture of machinery and equipment n.e.c., Manufacture of
motor vehicles, trailers and semi-trailers, Manufacture of other transport equipment

45
See Annex 1 for a more detailed description of the indicator

39
Medium low-technology manufacturing
19 Manufacture of coke and refined petroleum products
22 to 25 Manufacture of rubber and plastic products, Manufacture of other non-metallic mineral products,
Manufacture of basic metals, Manufacture of fabricated metal products except machinery and equipment
33 Repair and installation of machinery and equipment
Low-technology manufacturing
10 to 18 Manufacture of food products, beverages, tobacco products, textiles, wearing apparel, leather and related
products, wood and of products of wood, paper and paper products, printing and reproduction of recorded
media
31 to 32 Manufacture of furniture, other manufacturing

40

Economic development is linked to major changes in
the structure of economies. Changes in technology
and skills enable economies to produce the same
goods at higher levels of productivity, and to develop
new products and services. At the same time,
consumer demand and derived demand for
intermediate goods and services shift to different sets
of goods. This process of long-lasting changes in the
set of goods and services produced and in the
composition of capabilities the physical and
human capital base as part of the factors of
production is called structural change. This long-
term process should be distinguished from shorter-
term changes in the structure of economies that may,
for example, be induced by economic bubbles and
their collapse. In this regard, the chapter does not
focus on sector changes occurring during the current
financial crisis.
Structural change is not only related to changes in the
composition of economies associated with economic
development. The growth potential of economies is
also affected by the sectoral composition of output
and employment. Some sectors experience higher
long-term growth than others, leading to shifts in the
shares of these industries in the economy. However, it
is important to note that the structure of the economy
can also change with no positive impact on economic
growth, if structural change increases the share of
sectors with low growth potential. Thus, the structural
composition of economies and structural change are
important elements to be addressed by economic
policy making in order to ensure that the positive
growth enhancing structural change is facilitated. For
this reason, this chapter deals with the pattern of
structural change observed over recent decades. The
driving forces of structural change are technological
change and its impact on productivity, as well as
changes in the structure of demand associated with
changes in the prices of goods and aggregate income.
These determinants are interrelated and difficult to
disentangle, but they explain a large part of the
observed trends in structural change.
The primary aim of this chapter is to explore broad
patterns of structural change and its determinants with
a look at the policy relevance of structural change for
European policymaking. Europe is currently
experiencing an economic crisis the impact of which
on individual EU Member States varies. Within this
context, the importance of economic structure is
amplified when international trade and the sectoral
distribution of employment and output across the
production of tradable and non-tradable output are
taken into account. International trade can modify,
41
Chapter 2.
STRUCTURAL CHANGE
deepen and relax patterns of structural-change across
countries (McMillan and Roderick 2011).
The chapter covers the main trends of structural
change, the drivers, and the role of policies and
institutions in the process.
Broad trends of structural change are associated with
economic developments that are quite robust and
homogenous over time in the countries under
consideration. The share of agriculture is declining,
while the share of manufacturing displays a hump-
shaped pattern and the share of services is increasing
for almost all industrialized countries. The primary
drivers are productivity improvements based on
technical change and innovation, and changing
patterns in demand due to income effects and price
changes. International trade also has an important
influence on differences in economic structure across
countries.
Growth-enhancing structural change is associated
with the upgrading of capabilities, as well as with a
process of creative destruction. This process can be
observed by analysing a countrys export basket. In
more advanced economies, industries producing more
sophisticated and complex products are replacing
other industries. However, more sophisticated and
complex products require specific knowledge-bases
and specialisation patterns. The process of
reconfiguring capabilities and the range of products
produced by an economy is thus an important part of
the interaction between structural change and the
international division of labour.
The analysis of the relationship between broad
policies and institutions within the process of
structural change reveals that policies can guide
structural change, but that there are also important
limits to the impact of policies due to the existing
structure of economies. Because of international trade
and the associated specialisation patterns, economies
have different industrial structures. Therefore, policy
intervention should aim to support growth-enhancing
structural change by developing and building upon
existing strengths, rather than taking a completely
open approach.
2.1. BROAD TRENDS IN STRUCTURAL CHANGE
Structural change originates from microeconomic
changes which affect economic sectors in different
ways and with different magnitude. The changes at
the microeconomic level are important at the
aggregate level, because they are systematic and
affect the long-run performance of economies. The
result is that some sectors experience higher long-
 term growth rates than others, leading to shifts in the
shares of these industries in the aggregate. This
process unfolds over longer periods of time. This
chapter looks at the long-run changes between
sectors. These changes are best outlined using a
simple sectoral disaggregation which breaks down the
economic aggregates into three sectors:
1. Agriculture, fishery and forestry;
2. Industry in a broad sense covering manufacturing,
mining, construction and public utilities, and
3. Services, covering the different business, personal
and public service sectors.
In presenting the broad trends, results are given for
the manufacturing sector as a whole, as much
discussion of industrial policies focuses on the
manufacturing sector. In a subsection of the chapter,
services are examined in greater detail, as services
have become the dominant sector in terms of
employment and production in all advanced
economies.
Figure 2.1. Structural change in a historical perspective: Value-added shares for six European countries
Source: WIFO calculations based on data from University of Groningen and EU Klems
The broad trends in structural change are quite similar
across countries in the course of their economic
development. As economic development gets under
way, the share of agriculture in national employment
and value added falls, while there is a rapid increase
in the share of manufacturing and services. The
resource reallocation process associated with
structural change shifts economic activities from
agriculture to industry and services.
42
This pattern can be easily identified using historical
value-added shares for six European countries
(Belgium, Finland, France, Italy, Netherlands, Spain
and Sweden) over the 19th and 20th centuries.
Unfortunately, information is only available for these
six countries. Figure 2.1 plots the current value
added shares over the past 150 years. In order to
make the patterns comparable across countries,
nominal sectoral shares (nominal value added of a
sector in proportion to total nominal value added) are
plotted against the level of economic development
measured as the logarithm of per capita GDP in 1990
dollars (Bolt and van Zanden 2013). The line in the
figure is a polynomial prediction that does not take
into account country weights. The prediction provides
a better perspective of the association of the sectoral
value-added shares with economic development
(measured in GDP per capita at constant prices).
This historical pattern is quite similar to the pattern
identified in the cross section of a large number of
countries. Indeed, the historical patterns of structural
change would be of much less interest if findings for
the broad patterns of structural change were very
different for countries that are currently becoming
wealthier, in which case information on historical
patterns of structural change would not be very useful
for policymaking today. Figure 2.2 shows that the
patterns are quite similar if a very different dataset is
used, which has broader country coverage and
comparatively short time coverage. The National
Accounts Dataset collected by the United Nations
Statistics Division provides information on value
added shares for 164 countries over the period from
1960 to 2010.
Figure 2.2 plots the current sectoral value added
shares against GDP per capita from the UN National
Accounts dataset. The sectoral breakdown is again
agriculture, manufacturing, industry (manufacturing,
Figure 2.2. Structural change in the cross section: evidence from value added shares for 164 countries, 19602010
Source: WIFO Calculations based on national accounts statistics from the UN.
mining, utilities and construction) and services. The
line in the Figure 2.2 corresponds to a polynomial
prediction without country weights. The prediction
makes it possible to see the point estimate of the
association between sectoral value added shares and
economic development (measured in GDP per capita
in constant prices) which is independent of individual
countries.
Figure 2.2 confirms the basic regularities for
structural change found in the historical data.
However, the country coverage is much more
heterogeneous in terms of sector development. There
are countries in the sample that have a share of
agriculture of around 80% at very low levels of
economic development. A few countries even have
service shares as low as 10% of GDP. In addition,
there are countries that have very high value added
shares in manufacturing and industry. The results for
industry are often driven by countries with important
natural resources and a high share of mining in GDP,
such as oil-producing countries.
43
A visual inspection of the figures for manufacturing
shows that at a value of log real GDP per capita in
2005 of around 9, which corresponds to USD 8,100 in
2005 prices, the manufacturing share begins to
decline on average. This is broadly in line with the
evidence from the historical time series.
Figure 2.2 reports the results from a more rigorous
test using regression analysis. The sector share in
nominal value added is regressed on log real GDP per
capita (RGDP)46 using a fixed-effects regression in
order to control for unobserved, country-specific
factors affecting the composition of country shares.
Real per capita GDP of USD 8,100 (in 2005 prices) is
used to divide the country-year observations into two
sub-samples. For agriculture, it is observed that the
negative relationship between the agricultural share
and economic development is less strong for the sub-
sample covering the country-year observations with a
real GDP per capita above US$ 8100. This may be
related to the fact that for these countries, the value
added share of agriculture is already very small (4%
on average) for the more
46
Whilst Figure 2.2 displays a degree of non-linearity for certain
sectors, a linear relationship was modelled to illustrate the
direction and strength of the relationships in the two sub-
samples.
Table 2.1. Value-added share regressions for cross-section data, UN National Accounts data for 164 countries, 1960
2010
y<US$8100 y>=US$8100 y<US$8100 y>=US$8100
all observations sample sample all observations sample sample
Agriculture share Service share
RGDP 0.0869*** 0.117*** 0.0623*** 0.0543*** 0.0655*** 0.106***
(0.001) (0.002) (0.002) (0.002) (0.002) (0.007)

Observations 5872 4505 1367 5830 4463 1367

2
R 0.866 0.835 0.758 0.690 0.651 0.691
Manufacturing share Industry share
RGDP 0.00816*** 0.0257*** 0.0828*** 0.0343*** 0.0553*** -0.0436***
(0.001) (0.001) (0.003) (0.002) (0.002) (0.006)

Observations 5838 4471 1367 5872 4505 1367

R2 0.720 0.738
Source: WIFO Calculations
Note: Standard errors in parentheses. *** denotes significant estimate.
developed economies in the upper sub-sample while
it is still substantial (an average of 23%) for the sub-
sample covering the poorer countries. For the service
sector, an acceleration of the service share value-
added for the second sub-sample (country-year
observations with a real GDP per capita above USD
8100) can be seen. This finding corresponds to the
stylised facts reported by Buera and Kaboski (2012a)
for historical time series covering a larger set of
countries. Different relationships are found for the
manufacturing and industry shares in the two sub-
samples. There is a positive relationship for the first
sub-sample (real GDP per capita below US$ 8100)
and a negative relationship for the second sub-sample
(real GDP per capita above USD 8100) is observed.
The negative relationship between economic
development and the manufacturing share in the
second sub-sample is stronger than the negative
relationship between the industry share and economic
development for the same sample. This is partly due
to oil-exporting countries which have a high industry
share and a high level of real GDP per capita with a
low manufacturing share (e.g. Saudi Arabia, Kuwait,
United Arab Emirates) and partly due to the fact that
utilities and construction which are also part of the
industry share do not show a very strong association
with the level of economic development as services
and the manufacturing sector.
0.821 0.644 0.689 0.723
handicrafts while market services played even less of
a role. Successive industrial revolutions, exemplified
by the creation, diffusion and use of new
technologies, led to a gradual increase in productivity
in both the primary and secondary sectors. The share
of income spent on food decreased and employment
in the primary sector declined relative to the other
sectors. In manufacturing the increase in productivity
has led to lower prices. In the course of economic
development this has resulted in lower factor demand
(demand for labour, for instance) once productivity
outstrips the growth in demand for manufactured
goods. Over time, the tertiary sector has gained in
importance, both in terms of employment and output,
as enterprises have demanded support services and
consumption patterns have shifted towards services,
and productivity gains in manufacturing have become
much higher than those in service sectors.
Figure 2.3. Stylised broad patterns of structural
change

It is important to note that these patterns are not

restricted to nominal value-added shares but also
show up in employment shares. Figure 2.3 shows this
process in a stylised way. It shows that economic
Source: WIFO
development has consisted of a gradual shift from
agriculture to manufacturing and services, followed
by a shift from manufacturing towards the service
sector. In other words, when the market economy first
emerged, a vast majority of workers were employed
in agriculture, which accounted for the largest share
of production. The production of goods was limited to

44
2.2. PRODUCTIVITY IMPROVEMENTS AND CHANGES
IN DEMAND AS DRIVERS OF STRUCTURAL
CHANGE
The need to understand the mutual interdependency
between economic growth and structural change
requires analyses that enrich the general one-sector
macroeconomic perspective with multi-sectoral
perspectives, in order to understand the main
economic mechanisms that drive broad patterns of
structural change.
There are two central, possibly complementary,
explanations of the observed patterns of structural
change in the literature. The first explanation relates
to differential patterns of technical change between
different industries. The second explanation relates to
the different income elasticity of demand between
products of different sectors. From a conceptual
standpoint, the potential significance of these
mechanisms in explaining the broad trends of
structural change has long been recognized. For
instance, a pioneer in the theory of structural change,
Fourasti (1949), explained economic development
based on the combination of differential productivity
growth rates across agriculture, industry and services,
and the differential income elasticity of demand.
Technological progress was the main driving
mechanism behind structural change in his theory,
while the income elasticity of demand provides
something like a sorting mechanism that gave new
weights to the sectors. Fourasti maintained that in
the long run, the sorting mechanism of demand
dominates supply-side forces in shaping the economic
structure of countries. As income rises, the demand
for primary products will become saturated first,
followed by the growth of demand for manufactured
goods, which become eventually saturated, and an
increase in demand for products in the tertiary sectors
take place. Fourastis vision of the three-sector
hypothesis is one of the most elaborate theories of
structural change, but his explanation of why
products from industry can become saturated neglects
the role of intermediate inputs from industry that are
used in all sectors of the economy. A decline in the
demand for consumer goods in manufacturing does
not necessarily imply a declining share of the
secondary sector in total value added.
Other economists held such views at the time. For
example Kaldor (1981, 1996) argued that expanding
domestic and international markets engendered a
process of cumulative causation in which
manufacturing growth played a central role as many
growth-enhancing learning activities such as R&D
and the mechanisation of activities are closely related
to manufacturing. This allows a higher rate of
productivity growth in the manufacturing sector.
Today manufacturing is an important sector, but it is
also recognised that manufacturing industries are
heterogeneous, and that there are important
45
production and demand linkages which play a
significant role in the process of economic
development. More recently, the theoretical literature
has examined the conditions under which the two
determinants of different productivity growth (e.g.
Ngai and Pissarides 2007) and differential income
elasticities of demand (e.g. Echevarria 1997,
Kongesamut et al. 2001) can lead to an aggregate
balanced growth path. Herrendorf et al. (2013) claim
that the conditions under which these theories can
simultaneously generate balanced growth and
structural change are rather strict. Theories of
balanced growth, which constitute the workhorse of
growth theory, may not provide the right analytical
tools to explain the broad set of empirical regularities
of structural change.
2.2.1. Interaction of supply and demand factors
Pasinetti (1981, 1993) emphasised the importance of
the interaction of supply and demand side influences
in determining the outcome of the process of
structural change. Pasinetti stresses the influence of
income elasticity on the pattern of demand Engel's
law47 and technological progress as the main drivers
of structural change and long-term economic growth.
Hlzl and Reinstaller (2007) identify two
mechanisms linking the inter-industry and intra-
industry dynamics: sorting and selection. Sorting is
based on the idea that the industrial composition of
demand varies with income growth. This captures the
observation that the consumption of agricultural
products rises proportionally less than aggregate
income. Consumer preferences and the demand
derived by other firms for intermediate goods have an
impact on the relative growth patterns of sectors
within an economy. Selection in turn reflects price
competition within and between sectors: firms or
sectors able to produce the best value for money will
be able to increase their demand and grow faster.
Pure demand-side explanations of structural change
emphasise that changes in consumption associated
with income effects are a central driving force behind
the process of structural change. Rising income leads
to demand shifts from necessities towards
manufactured goods and then towards services (e.g.
Echevarria 1997, Kongsamut et al. 2001). However,
pure demand-side explanations do not take into
account the observed persistent differences in
technical change and productivity across sectors.
Baumols (1967) theory of imbalanced economic
growth is perhaps the most important supply-side
explanation for why the tertiary sector will gain in
importance over time. Baumol divides the economy
47
Strictly speaking, Engels law refers to the low income
elasticity of food but in the literature on structural change
Engels law is used to refer to structural change driven by
nonlinear income effects that affect demand for all types of
goods (e.g. Pasinetti 1981, 1993, or Iscan 2010).
into two types of activities: technologically Table 2.2. Dynamics in value added shares 1975 to 2005
progressive activities in which innovations, capital between EU-15 Member States and the US
accumulation, and economies of large scale all make
for a cumulative rise in output per man hour and Agriculture Industry Services
activities which, by their very nature, permit only EU-15 avg. value 2.3 26.2 71.5
sporadic increases in productivity (Baumol 1967, p. added share 2005
415-416). Baumol contends that productivity growth US avg. value added 1.8 23.8 74.4
in progressive activities drives wage growth in the share 2005
economy as a whole, causing relative costs in non-
progressive activities to rise. This leads to a fall in the
relative weight of the non-progressive sectors or, if EU-15 change in share 5.3 11.1 16.5
1975-2005
the relative outputs are maintained, to a slowing of
the aggregate growth rate, as an increasing proportion US change in share 3.7 7.8 11.5
of resources must be channelled into these activities. 1975-2005

Little is known about the relative importance of these EU-15 change in 1.6 1.9 1
two mechanisms in the process of structural change inequality 1975-2005
and economic development from the empirical
perspective. Part of the difficulty in understanding the US change in 5.6 2.1 1.4
inequality 1975-2005
relative importance of the supply-side and demand-
side drivers of aggregate growth is the paucity of data Source: WIFO calculations, EUKLEMS for the EU-15, Bureau of
Economic Analysis (BEA) for US states.
on the services sector. There are still significant Note: Differences in the industry classifications limit the
measurement issues with service sector output, value comparability of the data across regions (EU-15 and US). For
added and productivity growth. It is also very difficult European data NACE 1 is used, while US data follow the NACIS
to establish the importance of preference parameters classification. The change in shares between 1975 and 2005 refers
to the difference in the value added shares between the two
governing the income elasticity of demand for periods. Inequality is measured using the Gini coefficient within
services in a rigorous way. the broad regions (US, EU-15).

A closer look at the empirical broad patterns of
structural change reveals that both processes are
relevant. The gradual shift in value added and
employment shares from agricultural to
manufacturing and onwards to the services sector
seems to be mainly due to changes in market demand.
However, this account of the shift of demand also
needs to take into account big productivity
improvements in agriculture over the past decades.
Thus the mechanics of Engels law that as income
rises, the proportion of income spent on a good falls,
even if actual expenditure on it rises needs to be
complemented by an account of productivity
improvements. Up to now, there is no clear evidence
as to whether technological progress or changes in
demand is the driver of structural change. Baumol et
al. (1989) and more recently Nordhaus (2008)
provide empirical evidence favouring the
technological explanation. In contrast, Dietrich and
Krger (2010) find empirical evidence for the
demand story for the rise of the service sector in
Germany. Additionally, results by Curtis and Murthy
(1998), Rowthorn and Ramaswamy (1999) and
Peneder et al. (2003) suggest that the income
elasticity is greater than unity for most service
branches as well as for aggregate services, and below
unity for manufacturing branches.
Productivity and growth decompositions generally
lead to a view that structural components appear to be
largely dominated by the intra-industry and intra-firm
effects of productivity growth (Isaksson 2009). The
empirical literature confirms that the broad patterns of
structural change are driven by both demand-side and
supply-side dynamics. It confirms that structural
change can generate both positive and negative
contributions to aggregate productivity growth. If
structural change reallocates resources towards
sectors with higher potential of productivity growth,
structural change is growth-enhancing, and if
structural change shifts resources and employment
towards sectors with below-average productivity
gains, structural change may be growth-reducing. In
many cases the effects of structural change net out,
and structural change on average appears to have only
a weak impact on aggregate growth over short time
periods. Hence, if certain types of industries achieve
higher rates of productivity growth and

46
Box 2.1. A comparison between patterns of structural change in EU-15 Member States and US states
The fact that the majority of industrialized economies experienced a shift from manufacturing to services in recent
decades illustrates the similarity in the change in contributions of agriculture, manufacturing and service sectors to
total value added in the US and the EU. Between 1975 and 2005, the shares of services saw double-digit increases in
both the US and the EU-15. The increase of the services sector has taken place largely at the expense of industry
(mining, utilities, construction and manufacturing). On average across the US states, the share of services has
increased by 11.5 percentage points. This shift has been even more pronounced in the EU, where the share of services
increased by 16.5 points to 71.5 % in 2005. During this time, some member states (notably Greece, Spain and
Portugal) experienced a substantial catch-up. The share of manufacturing in the EU fell by 11.1 percentage points on
average. The decline has been slightly less (7.8 percentage points) in the US states. The comparatively small shares of
agriculture decreased further, more so in the EU than in the US.
One important question is whether structural change leads economies to become more similar over time or magnifies
regional and interregional disparities in the composition of aggregate output. This question is important for several
reasons. The tradability of agricultural and manufacturing commodities coupled with positive agglomeration effects
in their production foster regional specialisation, yet there are limits to specialisation in the production of some
services, which may not be tradable. Increasing structural disparities between the regions together with stark
differences in productivity developments across the three sectors have the potential to reduce or increase income
inequality within the regions. Economic policies aimed at attaining or maintaining a certain composition of output
also need to take into account regional inequalities and the underlying specialisation trends.
To answer this question in a simplified way, the inequality of the shares of agriculture, manufacturing and services
between the 15 EU Member States and the 50 US states relative to their aggregate economies (US and EU-15) is
considered. The preferred measure of structural cohesion is the difference between the values of the ubiquitous Gini
coefficient of inequality, calculated across the member states and federal states of both regions for the years 1975 and
2005 (Table 2.2).
Negative differences mean that regional (country/state) differences have decreased. This is the case for the share of
services, as the nationwide rise in the contribution of services has been due in part to their low tradability and local
character. On the contrary, as expected, disparities in the structure of manufacturing have increased.
An apparent difference between the 15 Member States and 50 US states lies in the evolution of the inequality in the
contribution of agriculture. Regional inequality has increased in the EU-15, while it has decreased in the US. This
may be related to the fact that the US has had a common agricultural market since its early days, with regional
differences and specialisation in agriculture taking place long before 1975. In the EU this process started around this
time. This may explain why differences appear to have decreased in the US, while they have increased in the EU.
Interestingly, quite similar patterns in inequality are observed for the industry and services sectors for the US states
and the EU-15. Industry shares have become more unequal across Member States and US states, and services shares
(almost by nature) have become more similar across the two regions.
expansion in output than others, structural change in services in the same way. In fact, the rise of the
favour of specific industries might still be conducive services sector taken as a whole has mainly been due
to economic growth. However, this might not be seen to the expansion of business services and some non-
at the aggregate level. The comparison between market services.
patterns of structural change in the EU-15 and in the
US shown in Box 2.1 confirms that structural change For a long time, the shift towards services was seen as
has been quite similar. Both have experienced a growth-reducing structural change. The rates of
dramatic growth in the value added share of services productivity growth in manufacturing and services
across all constituent EU countries and US states, and are very different and can to some extent explain the
this shift towards services has been associated with a large-scale labour reallocation in favour of the
relative decline in industry and agriculture. services sector. However, more recent economic
research clearly shows that many knowledge-
2.3. THE EXPANSION OF THE SERVICES SECTOR intensive services are important factors in economic
growth. For example, Pugno (2006) emphasises the
The analysis of broad changes has revealed that the importance of education and human capital formation
biggest shift experienced by industrialised countries for economic growth.
over the past decade has been the reallocation of
resources and employment linked to the growth in Buera and Kaboski (2012b) emphasise the skill
services. All highly industrialised countries have intensity of many service sectors and propose a
become service economies, in terms of the share of theory of the rise of the service economy based on an
value-added generated in the services sector and increasing importance of specialised highly-skilled
when employment shares are considered. This labour at high levels of productivity. Thus the rise of
structural change is uneven as it has not affected all

47
the service economy is a growth in the range of
services that are market-produced relative to those
that are home-produced. Buera and Kaboski (2012b)
provide an explanation of the rising level of skills and
skill premium that goes in hand with the rising
relative level of prices for services, which is
associated with changes in demand towards
knowledge-intensive services in the process of
economic development. In particular, the application
of modern information and communication
technology to the production of services has changed
the perception of services as low productivity and low
skill sectors of the economy. Eichengreen and Gupta
(2013) and Jorgenson and Timmer (2011) show that
more traditional services like lodging, housecleaning,
distribution, education and healthcare are increasingly
complemented by modern services such as banking,
insurance, communication and business services.
Figure 2.4 provides evidence on the heterogeneity of
services sector expansion in the most advanced
economies, including most EU Member States. In
these figures EUKLEMS data are used and four
different types of services are distinguished:
- Distribution
- Personal services
- Business services
- Non-market services (education, health and
government services)
Figure 2.4 reports both the value-added and the
hours-worked shares. While the share in hours
worked is almost constant, the value share of
distribution decreases with economic development.
The expansion of business services is more dynamic
in terms of valued added shares than for hours-
worked shares. However, the opposite seems to be
true for non-market services. This shows that the
expansion of the service share is mainly driven by the
expansion of two quite different service subsectors:
business services and non-market services. While the
pattern of expansion of non-market services
government services, health and education to name
the most important could be explained by a supply-
side cost disease argument, the same argument does
not apply for the expansion of business services,
because for many countries the increase in economic
weight is more substantial in terms of value added
than hours worked.
This evidence shows that services are heterogeneous
and is compatible with the argument provided by
Peneder et al. (2001) and Buera and Kaboski (2012b),
indicating that the rise of the service economy has
primarily been driven by the growth of knowledge-
based services.
Even if the shift in structure towards services has
reached unprecedented proportions, the understanding
of the factors accounting for the shift to services is
still partially contested. This is related to a number of
48
issues. Different mechanisms have been proposed to
explain the shift of economic activities towards
services. The thesis of marketization or de-
marketisation of home production proposed by Buera
and Kaboski (2012a) is one that combines the
differential development of technology with a
mechanism of a shift in demand.
Schettkatt and Yocarini (2006) emphasise the
importance of demand-side explanations. They argue
that shifts in demand associated with income effects
have been the driving force of the expansion of
services employment in past decades. However, the
different productivity developments between services
and manufacturing are also important. Price trends in
some services support this view (e.g. Schettkat and
Yocarini 2006): prices of services generally rise more
than prices for manufactured output. However, as
emphasised by Peneder (2001) some services sectors
are obviously technologically progressive. Jorgenson
and Timmer (2012) clearly show that price and
productivity developments in distribution sectors are
very different from other service sectors.
Table 2.3 gives an indication by using the relative
price development of sectoral prices compared to the
GDP deflator as a measure of sectoral price
developments. Values below 1 indicate that price
developments were below the aggregate price
development (GDP deflator). Conversely, a value
above 1 indicates that prices rose faster than average.
The table displays average values for the EU27
Member States and the associated standard
deviations. Across Member States, agriculture and
manufacturing have had a below-average price
development. The price development in distribution
was on average approximately the same as for
aggregate prices. For personal services, business
services and non-market services, an above-average
price development is observed. These price trends are
consistent with the view of differential productivity
developments across services and manufacturing and
higher productivity dynamics in manufacturing.48 The
associated standard deviations show that these
differences are statistically significant. Nevertheless,
it is also important to note that these price series are
themselves subject to a considerable composition
bias, as it is very unlikely that the structure of these
quite aggregate sectors remained identical over
time.49
48
Price developments of manufacturing products would be
further below the aggregate price development if they were
corrected for increases in the quality of finished products (see
Cummins and Violante (2002).
49
This remark is important for the comparison of real shares over
time. Structural change is a process that changes the weights of
economic activities in the aggregate. Moreover, structural
change is driven by differences in demand and productivity
that react to or determine prices. Therefore these data not only
identify a price effect but also a quantity effect associated with
the changing weights of economic activities.
Figure 2.4. Service shares and economic development

(a) Value added shares

(b) Shares in hours worked

Source: WIFO calculations, EUKLEMS

49
Box 2.2. Household production and structural change

It has been widely recognised that income and wealth generated by household production could introduce a bias in
the measurement of the economic structure. Kuznets (1944) and Clark (1958) already indicated that the neglect of
home production leads to a significant underestimation of national income in general and the contribution of
agriculture, construction, and services to national income in particular. These missing activities contribute to
economic welfare and can be marketised to different degrees across countries, thereby affecting the measurement of
sector shares, since only market services are taken into account in the official statistics.

Hill (1977) defined household production as economically productive households or do-it-yourself activities that can
be provided through the market by choice. Home-produced and market-produced services are gross substitutes.
Cooking and cleaning are productive activities because the market can provide them, whereas eating and sleeping are
non-productive activities because the market cannot provide them. Unfortunately, no systematic data are available
which would allow differences in structural change to be quantified across EU Member States. The results for the US
suggest that incorporating the value of non-market home production increases the level of nominal GDP. Bridgman
et al. (2012) estimate that in 1965 household production increased GDP by 39%, and by 26% in 2010. The relative
decline in home production is almost solely due to the reduction of hours spent in home production by women. Their
contribution declined from 40 to 26 hours during this period, whereas the number of hours spent by men increased
from 14 to 17 hours. Thus the allocation of hours between market and home production can also be considered in the
context of structural change. Yet Freeman and Schettkat (2005) and Rogerson (2008) suggest that it is not a driver of
structural change but part of the increase in leisure deriving from the reallocation of labour from home production to
the market production of services.

It is also important to take into account the impact of technical change on home production. A good example is the
contribution of the mechanisation of home production, namely the growth of the manufacturing of household
appliances. The spread of washing machines, dryers, vacuum cleaners, microwaves, and other home appliances was
accompanied by declines in domestic servants, laundries and drycleaners. Buera and Kaboski (2012a) emphasise that
innovations which change the scale economies of productive activities are an important determinant of the boundaries
between home and market production of goods and services over time. They argue that scale economies are a driving
force in the process of marketisation of services, but that this process can also be reversed if technological change and
mechanisation lowers the cost of producing services and scale economies, as households value the flexibility of home
production. Their result suggests that the spread of manufactured goods into the home leads to a demarketisation of
services and a growth of manufacturing relative to services. Technical change that leads to an increase in the
economies of scale of services will lead to the marketisation and relative growth of the service sector. However, it is
important to note that the empirical verdict on the importance of the thesis of the marketization and demarketisation
of home production for structural change is still not settled.

It should also be noted that the de-marketization of certain activities, because they can be done at home at a low
cost in terms of time, can potentially increase labour supply and reinforce the tertiarization of the economy
because services are labour-intensive.

2.3.1. Interaction of manufacturing and services
Another important explanation that has been brought
forward is the hypothesis of the inter-industry
division of labour. It is sometimes claimed that
outsourcing jobs from manufacturing to services is a
primary driver of the rise of service sectors (see
Schettkatt and Yocarini (2006) for a discussion).
However, most of the studies using input-output
analysis come to the conclusion that outsourcing from
manufacturing to services took place at a very modest
rate50. According to Gregory and Russo (2004), the
rise of business services is largely explained by
outsourcing from other service sectors. This shows
that the trend towards an increasing services share
cannot be understood without considering changes at
the microeconomic level. The interaction between
manufacturing and services has become more
complex. Services and manufactured goods are used
as intermediary inputs to produce a larger number of
final products (goods and services).

50
The methodological limitations of using input-output analysis
to examine the outsourcing process are discussed in Montesor
and Vitucci (2007).

50
Table 2.3. Relative price developments at the sector
smaller than the service content of manufacturing and
level, 19952007 has increased much less over time.
Average Standard deviation These results explain that there is a high degree of
Agriculture 0.77 0.18 complementarity between manufacturing goods and
services, but that this complementarity is biased
Manufacturing 0.82 0.13
towards the increasing importance of services as
Construction 1.22 0.20 inputs to manufacturing. Services such as
Distribution 0.97 0.09
maintenance and training are very important elements
in the delivery of complex manufactured products. At
Personal services 1.17 0.16 the same time the importance of specialised services
Business services 1.13 0.13 such as financial intermediation, communications,
insurance and knowledge-intensive business services
Non-market services 1.18 0.16 (KIBS) are becoming important inputs in the
Source: EUKLEMS production of sophisticated manufacturing output.
This process is one of several explanations for the
The increasing contribution of the service industry at increasing contribution of services to the overall
the expense of manufacturing can in part be explained output of an economy. Figure 2.5 shows that an
by an increasing service content of manufacturing increasing share of intermediate services used in
final output, reflecting the total value of the services modern manufacturing is being imported, reflecting
required for the development, production and the fragmentation of production processes within and
marketing of a modern manufacturing product. The across national borders which was made possible by
service content of manufacturing has been growing in the ICT revolution (Guerrieri and Meliciani 2005).
the EU and elsewhere in the world. In 2011, the share The relocation of business processes from one
of value added embodied in manufacturing final country to another also affects the structure of an
output that was created in the service industry ranged economy and its services share. However, the degree
from above 40% in Belgium, Ireland and of offshoring and outsourcing is dependent on the
Luxembourg to below 30% in Romania, United particular industry and activity.
Kingdom and Greece (see Figure 2.5). This means
that more than a third of the value of a European In the discussion of the rise of services, two issues
manufacturing product that is sold to final users is have so far been neglected. The first issue relates to
created in the services sector. Whereas manufacturing the quality of data on services. There are still
products are also used for producing services, the measurement issues associated with services sector
manufacturing content of services is about three times
Figure 2.5. Value added decomposition for manufacturing production in per cent, 2011

Note: Countries are ranked according to their domestic value added share (i.e. Domestic non-services + domestic services)
Source: WIOD, wiiw calculations

51
output, value added and productivity.51 Griliches
(1992), for example, documented that services sectors
with hard-to-measure outputs, such as health services,
experienced the largest labour productivity slowdown
after 1973. In many services there are
economic growth but also the expansion of some non-
market activities such as education. In this case the
productivity and quality of the educational system are
of primary importance. The productivity of the
education system needs to be measured in terms of its
quality in providing the right competences and

Figure 2.6. Share of value added created in non-services sectors embodied in services production (%)

Note: Services sectors are NACE Rev. 1 50 to 95.

Source: WIOD, wiiw calculations

conceptual and empirical problems in measuring capabilities.

output and prices. Therefore, there are still substantial
differences in the measurement of productivity in 2.4. HETEROGENEITY OF STRUCTURAL CHANGE IN
manufacturing and in services, and thus major limits EUROPE
in the comparability of its growth rates across these
sectors. These problems primarily affect the value The discussion so far has shown that broad trends in
added shares in GDP and the identification of structural change are quite homogeneous across
productivity developments, while employment shares countries.
are unaffected by these issues. However, the
Table 2.4 summarises developments between 1995
importance of human capital and education for
and 2011 for the EU27 countries. There has been a
economic growth in developed economies (e.g. Lucas
decline in most production sectors, both in terms of
1988, Pugno 2006) leaves an additional question
employment shares and nominal value-added shares.
mark concerning the usefulness of comparing sectoral
The biggest reductions are for agriculture and the
productivity differences. If human capital is essential
manufacturing sector. In the period 2005 to 2011,
for economic growth and the educational sector
construction also shows a negative trend for
provides most of the human capital used in the form
employment as well as value- added shares, reflecting
of capabilities in the manufacturing and services
its sensitivity to aggregate downturns (cf. Hlzl et al.
sectors to develop new products and to improve
2011). These declines have been offset primarily
productivity, then the reallocation of resources to a
through the expansion in business services and non-
low-productivity activity such as education may be
market services. Overall, these results are consistent
the reason why productivity improves in other sectors
with a view suggesting that labour productivity
of the economy. Pugno (2006) explicitly takes this
growth has been especially strong in agriculture,
situation into account, concluding that not only can
manufacturing, distribution, mining and utilities.
the expansion of business services support long-term
Of greater interest is the development in the
51
dispersion within the EU27.
There is a literature on problems with measuring productivity
in services. One of the best overviews on measurement
problems in the services sector in general is provided by
Table 2.4 provides descriptive statistics of this
Triplett and Bosworth (2004). Diewert, Fixler and Zieschang dispersion, in terms of Gini indices. The Gini index is
(2012) cover banking services and Diewert (2011) cover a widely used measure of relative inequality. The
public services. coefficient takes values between zero and one, or, as

52
in Table 2.4, between 0 and 100 on the percentile
scale. A value of 0 corresponds to total equality. This
would be the case if all sectors had equal shares. The
higher the value of the Gini coefficient, the more
unequal the distribution. A value of 100 expresses
maximum inequality, for example if all countries had
different sectors. The formula for computing the Gini
coefficient is:
level of individual countries important differences
, can be observed.
where y_i are the sector shares sorted in an ascending
order and n the number of sectors.
Table 2.4 shows that the disparity is highest for
agriculture and for mining and utilities. The lowest
inequality ratings are for construction, distribution
and non-market services for the employment share,
and for distribution, non-market services and business
services for the value-added share. The changes in
inequality show that most services sectors
experienced a reduction in inequality during the
longer period 1995 to 2005. For the shorter time
horizon inequality increased for personal services,
probably due to a transitory divergence in the
consumption of personal services across countries.
However, for manufacturing we observe a rising
disparity across the EU27, for both the employment
and value-added shares.
Table 2.5 displays the heterogeneity of structural
change for the EU-15 over the longer time period
between 1975 and 2005 using EU KLEMS data. The
results confirm the earlier picture of the EU-15
experiencing a reduction in disparity of services
shares, along with an increase in disparity for the
production sectors, especially manufacturing both in
terms of employment and the value added share.
Overall these results suggest that there was
considerable heterogeneity in economic development.
The trends of structural change are quite similar
across countries, but we observe some divergent
developments, especially for the manufacturing
sector. The results in Box 2.2 (US EU Comparison)
also show an increasing inequality between the value-
added shares in manufacturing across Member States
and US states. The similarity in broad trends of
economic development is also compatible with
important heterogeneity at the country level. Further
analysis for the EU Member States shows that
common trends in structural change are able to
explain a large part of the development of
employment and value-added shares over time, but
also that idiosyncratic and contrasting elements are
important in determining the development of sectoral
shares across countries.
53
The same patterns of structural change emerge from a
more detailed look at structural adjustments in the
EU-1252 economies during the transformation from a
planned economy to a market economy, and then
during the process of integration into the Single
Market (see Box 2.3). The broad trend of a decrease
in production activities, especially manufacturing and
agriculture, and an increase in services sectors was
quite uniform across these countries. However, at the
2.5. INTERNATIONAL TRADE AND SPECIALISATION
AS DRIVERS OF DIFFERENCES IN THE
ECONOMIC STRUCTURE
What explains these differences in economic structure
across countries? The differences for the
manufacturing sector are the most interesting,
showing a polarisation process with an increasing
disparity across Member States. One important factor
is international trade. Openness and international
specialisation patterns clearly have an impact on
observed structural change. Connolly and Yi (2008)
claim that up to 30% of South Koreas catch-up
between 1962 and 1995 can be traced to its openness.
Matsuyama (2009) and Yi and Zhang (2010) show
that differences in the structure of economies can be
related to differences in international trade,
specialisation, and differences in economic
development, which is partly path-dependent.
Two important drivers of specialisation and structural
change are innovation, i.e. the creation of new
varieties of products, and the selection of new
products through the process of market competition
or changes in demand that affect the economic weight
of products and may even lead to the replacement of
products. The replacement mechanism is very
important as it captures the key mechanism behind
Schumpeters vision of economic development driven
by the process of creative destruction. This
perspective of qualitative change is closely related to
a view of economic development as a process of
structural change, where resources are continuously
reallocated from activities with low productivity to
activities with higher productivity.
This view has been emphasised in a series of
contributions by Hidalgo et al. (2007), Hidalgo and
Hausmann (2009) and Felipe et al. (2012) see also
Reinstaller et al. (2012) which linked the process of
economic development of a country to the idea of
changes in the space of its exported products. In such
a perspective, the overall complexity and
sophistication of a countrys productive structure is
52
Bulgaria, Cyprus, Czech Republic, Estonia, Hungary, Latvia,
Lithuania, Malta, Poland, Romania, Slovakia, Slovenia.
Table 2.4. Dynamics and heterogeneity of structural change in the EU27, 19952011
Employment shares

Share Change 1995 Change 2005 Inequality Change 1995 Change 2005
2011 2005 2011 2011 2005 2011

Agriculture 5.8 2.6 -0.6 39.6 2.9 1.1

Mining and Utilities 1.3 0.5 0.0 31.8 2.1 0.2

Manufacturing 15.3 3.2 1.9 16.6 4.0 1.4

Construction 7.1 0.8 0.7 9.6 2.2 5.0

Distribution 26.6 1.1 0.5 7.5 0.3 1.5

Personal Services 6.1 0.5 0.3 18.3 0.9 1.1

Business Service 14.6 3.0 1.7 18.9 1.8 1.4

Non-market Services 23.1 1.0 0.7 10.1 1.5 0.6

Value added shares

Share Change 1995- Change 2005- Inequality Change 1995- Change 2005-
2011 2005 2011 2011 2005 2011

Agriculture 2.7 2.2 0.1 30.3 1.5 0.9

Mining and Utilities 3.7 0.5 0.5 26 2.4 1.7

Manufacturing 16.8 3.0 0.8 18.2 5.2 1.7

Construction 5.6 0.6 0.9 14.4 2.8 1.1

Distribution 23.3 0.9 1.1 8.9 2.0 1.1

Personal Services 4.3 0.4 0.4 20 0.1 8.1

Business Service 25.3 3.3 1.3 11.7 0.3 1.4

Non-market Services 18.3 0.6 0.6 10.3 1.2 0.5

Source: WIFO calculations, Eurostat, National Accounts.
Note: unweighted averages; inequality is measured using the Gini Index across countries.

Table 2.5. Heterogeneity of structural change in the EU-15, 19752005

Employment share Value added share

Change in
Change in inequality
Inequality 2005 Inequality 2005 inequality 1975-
1975-2005
2005

Agriculture 38.5 1.8 29.4 0.6

Mining and Utilities 32.0 9.2 24.3 1.0

Manufacturing 15.2 7.6 16.5 7.8

Construction 14.6 7.0 13.3 5.4

Distribution Services 6.0 1.7 10.0 3.9

Personal Services 17.2 2.3 11.9 3.1

Business Services 20.3 2.4 13.1 2.7

Non-Market Services 10.7 5.8 9.8 3.5

Source: WIFO calculations, EU KLEMS

54
the key indicator to explain its economic The processes of variety creation and creative
development. destruction can be made visible by the appearance
and disappearance of exported products or product
Different abilities to accumulate capabilities to
classes. Exporting new products changes the
produce new improved products can explain
composition of the product basket of countries.
differences in their performance. This literature
Therefore the structural change in one country may
provides a novel way to study the differences in
affect the economic structure in other countries. The
structural change across countries.53
analysis of changes in the composition of export
Box 2.3. Structural change in the EU-12 during transformation and integration into the Common
Market

In general, the Member States which joined the EU in 2004 and 2007 had oversized and inefficient industrial sectors.
At the start of the transformation process, the high degree of industrialisation was a drawback. It implied, among
other problems, the underdevelopment of important services sectors. Due to comparative disadvantage, industry in all
the former communist countries suffered disproportionately from the 'transformational recession in the early 1990s.
The relative decline of industry went hand in hand with a rapid expansion of services sectors. By 2011, only the
Czech Republic and Romania had a manufacturing sector with a share in GDP of more than 20% about the same as
in two of the more industrialised older Member States: Germany and Ireland. In Hungary, Poland, Romania, and the
Baltic states, manufacturing industry managed to retain at least part of its previous position, thanks largely to active
restructuring and privatisation efforts, fostered in particular by FDI inflows. At the beginning of the 2010s, the shares
of manufacturing to GDP in the majority of EU-12 Member States were higher than in EU-15 economies. However,
this is in line with many developing economies.

The changes in employment shares in the EU-12 countries was even more dramatic during the last two decades.
Employment declined more than output and millions of jobs were lost during the transition from central planning to
market economies. Nevertheless, the manufacturing sector remains an important job provider, with the highest
employment shares in the manufacturing industry recorded in the Czech Republic, Slovakia and Slovenia. With the
exception of Latvia, Cyprus and Malta, manufacturing accounts for more than 15% of total employment in all EU-12
Member States. Similarly high shares of manufacturing employment are recorded for only a few EU-15 countries:
Portugal, Italy, Austria, Germany and Finland.

Structural change has been more pronounced in Bulgaria, Romania, Latvia and Lithuania than in the Czech Republic,
Hungary, Estonia or Poland. Furthermore, the earlier transition period 1995-2000 was more profound than the
integration period immediately before EU accession (2000-2005). The most recent period, 2005-2011 is characterised
in several countries by more restructuring than before EU accession (for instance in the Czech Republic, Slovakia
and Slovenia). This period was also affected by the recent economic crisis which hit manufacturing, construction and
tradable services much harder than other economic sectors. Among the EU-15, Sweden, Austria and Germany
experienced only small adjustments, whereas structural adjustments were more pronounced in Ireland and Finland.

Despite varying country-specific restructuring patterns, several stylised facts common to most countries can be
observed for the EU-12: the output shares of agriculture and manufacturing have declined whereas those of real
estate, renting and business activities, information and communication, financial and insurance services, as well as
public administration have increased. However, it must be said that the patterns of structural change were quite
different across individual countries. It is especially interesting that a number of new distinct features of structural
adjustment emerged during the relatively short crisis period between 2008 and 2011. Apart from a certain revival of
manufacturing (Hungary, Romania and the Baltic states) it was construction and trade which suffered most from
declining value-added shares during the crisis in a number of EU-12 countries. Structural adjustments were less
pronounced in the Czech Republic during this period (as in a number of EU-15 Member States, such as Austria,
France, Germany, Belgium, Italy and Sweden). In Poland the only EU Member State which did not experience a
decline in GDP during the crisis period a certain return to a traditional structural pattern occurred as a number of
productive sectors (energy, construction and trade) managed to increase their shares in GDP while the shares of
information, communication services and especially financial services showed some declines.

baskets of countries using trade data at the four-digit

53
Some recent contributions argue that the hump-shaped pattern
of development observed for the shares in value added of the
manufacturing sector over time can only be explained by
taking an open economy perspective and in the context of
international specialisation in trade (Reyes-Heroles 2012; Uy
et al. 2012).
product level for 232 countries covering the years
1995 to 2010 (cf. Gaulier and Zignago 2010) allows
the study of structural change and what is termed
creative destruction. We use the product space
indicators proposed by (Hidalgo et al. 2007; Hidalgo
and Hausmann 2009) that capture trade specialisation,

55
product complexity, and appearance and
disappearance of traded products across countries.
Figure 2.7. Average product complexity, density and co-appearance across income levels. Predicted values on the
basis of fitted fractional polynomials, 2010
Source: WIFO calculations; BACI dataset (Gaulier and Zignago 2010);
Figure 2.7 provides a summary of these indicators by
relating them to per capita income levels of the
countries. The upper panels (product complexity and
trade specialisation) show values for 2010; the lower
panels (change of trade specialisation and co-
appearance and disappearance of products) show
differences between 1995 and 2010.
The product complexity score (PCS) is shown in the
upper left panel. This indicator can be interpreted as
capturing latent information on both the depth
(capability to produce exclusive products due to high
levels of accumulated knowledge) and the breadth of
the knowledge base (capability to make many
products with different knowledge bases) needed to
be active in a specific product class (cf. Reinstaller et
al. 2012).54 It is constructed using information on how
many countries produce a specific product and on
how diversified these countries are. The plot in the
figure shows that more developed countries produce
more sophisticated products which require higher
54
The empirical range of the product complexity scores lies
between 4 and 2. These figures correspond to standard
deviations from the mean product complexity score normalised
to zero. Hence an indicator value of 2 indicates that the
complexity score of a product or product class is two standard
deviations away from the mean. A product with a complexity
score of zero indicates that relative to the entire sample it
has just average complexity.
56
capabilities and suggests that economic development
goes along with a perpetual structural change in the
export basket towards more complex products.
The upper right panel shows the product
neighbourhood density55. This indicator is a proxy for
the trade specialisation of countries. It exploits the
fact that similar products are related to each other by
drawing on common knowledge bases and similar
factors of production. It is therefore also a measure
for the factor substitutability across products. Higher
scores imply a higher specialisation. In order to plot
this indicator it has been averaged over products in
the product basket of a country. The plot shows that
countries at higher levels of economic development
tend to become more specialised in their exports. The
products they export are more closely related to each
other in terms of similar factor input requirements.
The lower left panel of Figure 2.7 shows the change
of the neighbourhood density between 1995 and
2010. It is plotted against GDP per capita levels in
2010. The figure suggests that trade specialisation
seems to be a fast process at lower levels of economic
development, while it starts to slow down at a GDP
per capita corresponding to about USD 3000 (or e8),
in a hump-shaped relationship.
55
At the product level, the indicator takes on values between 0
and 1, where 0 indicates no relation and 1 a perfect relation of
a product to the productive structures of a country the
specialisation pattern of a country.
The lower right panel plots the co-appearance index
of the products countries export. This measure
captures the presence of temporal clustering, where
products appear in the export basket across countries
at the same time. A higher value indicates a temporal
clustering and shows whether countries start
exporting these products simultaneously. The plot
shows that the co-appearance index follows an
inverted U-shape over levels of per capita income,
with a predicted maximum at an income level
corresponding to about USD 8000 per capita (=e9).
While a more thorough examination of the
relationship between the co-appearance index and the
change in neighbourhood density (specialisation) is
necessary, a first glance at the results suggests that
the co-appearance of products is closely related to the
dynamics of related specialisation. Increases in
specialisation appear to go along with bursts in the
Figure 2.8. Specialisation of productive structures and changes over the 19952010 period across income levels,
predicted values on the basis of fitted fractional polynomials
Source: WIFO calculations; BACI dataset (Gaulier and Zignago 2010)
export activity of products.
Figure 2.8 displays the export share weighted
displacement index. This index measures the number
of disappearances of some product classes within a
specified time window after a country has started
exporting another product.56 It captures the creative
destruction induced in a productive system when a
country starts making a specific product. Negative
values indicate that a country exports mostly products
56
The variable takes on values between -1 and 1. See Klimek et
al. (2012) for details on the indicator.
57
which across all countries tend to be displaced by
other products. Positive values indicate that products
which tend to displace other products have a higher
weight in the export basket of a country. Such
products can be considered to be more innovative,
high-end products. Beyond a threshold income level
close to USD 3,000 per capita, displacing products
start to dominate the export basket of countries. This
figure gives a clear indication that the characteristics
of traded products change in countries with levels of
income per capita in the range between USD 3,000
and USD 8,000.
As Figure 2.9 shows, the export baskets of the EU-27
consist by and large of products with positive
displacement scores. This indicates that most of the
exports of the Member States are products from the
upper end of the quality ladder. The data also shows
that inside this group of products with positive
displacement scores, some Member States have a
higher export share in products with above-average
complexity scores whereas others have a higher share
of products with average or below-average
complexity scores.57 The product complexity score
can be taken as a measure that captures the difficulty
of imitating exported products. For Member States
with average or below-average complexity scores,
57
Unreported results show that the export basket of catching-up
countries such as Brazil is dominated by products with
negative displacement scores and a relatively high share of
products with below-average complexity scores inside this
product class.
this evidence implies that they produce up-market
products that are easier to imitate. As a consequence,
they are also subject to more intense price
competition from lower-income countries than
Member States producing innovative products that
rely on a more complex knowledge base and
therefore are also more difficult to imitate.
Further results show that across manufacturing
sectors, product classes with negative and positive
displacement indices co-exist. In the chemical
industry for instance, the share of the two product
categories is almost equal. By contrast, in the
machinery and equipment industry the share of
displacing products, i.e. products with a positive
displacement index, outweighs the number of
products that tend to be displaced, whereas the
opposite situation exists in the textile and apparel
industries. Sectors thus undergo a permanent
restructuring process which is driven by changes at
the level of products or product classes. These results
show that more sophisticated products both in terms
of complexity and displacement scores are more
frequent in medium-high and high-tech industries.
Therefore, in the more advanced economies, sectors
producing more sophisticated products drive out other
sectors. However, it is important to stress from a
policy perspective the considerable path dependence
in the development paths of the productive structures
of economies (e.g. Reinstaller et al. 2012). This
implies that the diversification into economic
structures characterised by innovative products which
are difficult to imitate is harder to achieve by
countries lacking specific knowledge bases and
specialisation patterns than by countries that have
these capabilities. Thus an important limit to the
change in the export basket of countries is the path
dependency of industrial structure.
The path dependency of industrial structure also
suggests that it might be easier to lose some products
and competences in the process of international
competition than to build up different capabilities
which allow the differentiation of the product space
of a country. This dynamic process of reconfiguring
capabilities, competencies and the national product
space is part of the interaction between manufacturing
share and international competiveness.
While it is true that for most countries agricultural
and manufactured goods are the most important
tradables, the discussion should not be reduced to the
size of the manufacturing share alone. The
composition of the manufacturing share itself,
whether manufacturing consists of sophisticated and
complex products with unique features or mainly of
products which compete with goods from many
countries, is very important as this determines the
long-run position of countries.
58
2.5.1. The role of institutions in structural
change
The product space literature suggests that capabilities
are crucial in explaining differences in structural
change and economic development. On the other
hand, economic development is closely linked to the
institutional quality of countries. For example, Knack
and Keefer (1995) and Dollar and Kraay (2001) argue
that rule of law is an important driver of economic
growth.
The literature on institutions and economic
development suggests that many institutional
indicators are highly correlated with economic
development (e.g. Langbein and Knack 2010). A few
studies have provided evidence of causality running
from institutions to economic performance
(Acemoglu et al. 2001, Rodrick et al. 2004).
Reinstaller et al. (2012) confirm that product space
indicators capturing the complexity of the export
basket are closely correlated with institutional quality
and high knowledge intensity. However, in this
literature the relationship between structural change
and institutional quality is not made very explicit, as
it is not possible to measure structural change in an
unambiguous way. The problem is that structural
change can be growth-enhancing and growth-
reducing. In the presence of international trade the
reallocation of resources (e.g. labour and capital) can
lean towards high-productivity sectors or in the
opposite direction. Latin America has been cited as an
example of a larger region which in the past
experienced growth-reducing structural change. In the
1960s and 1970s in particular, economic policy
driven by macroeconomic populism and protectionist
import-substitution policies provided the basis for this
outcome (e.g. McMillan and Rodrick 2011). This
suggests that two different types of institutions and
policies are central to fostering growth-enhancing
structural change: institutions and policies to promote
the efficient reallocation of resources across sectors
and institutions, and policies to encourage the
development of capabilities which allow enterprises
to innovate.
The literature on market frictions in structural change
emphasises that aggregate outcomes not only depend
on rationalisation and reorganisation processes within
firms and industries, but also on the reallocation of
resources across sectors. Restuccia and Rogerson
(2013) survey the evidence and show that structural
change can be limited by the existence of regulations
and other frictions that inhibit the reallocation of
resources across sectors and firms. This can be costly
in a static sense, as the resources
Figure 2.9. Composition of the export basket in terms of product complexity and displacement indices: shares in total exports, EU27

AUT BGR BLX CYP CZE DEU

0.6369 0.6354
0.6 0.5619 0.4316 0.5302
0.5 0.4418
0.4 0.2990
0.2953
0.2408
0.3 0.2341 0.2299
0.1545 0.1866 0.1962
0.2 0.1344 0.1094 0.1209
0.0743 0.0922 0.0809 0.0551 0.0426
0.0639
0.1 0.0508
0
DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0
DI<0 DI>=0

DNK ESP EST FIN GBR

FRA
0.5 0.4507 0.4711 0.4748
0.45 0.4828 0.4804 0.5013
0.4
0.35 0.3163
0.2820 0.3256 0.3131 0.3142 0.2595
0.3
0.25
0.2 0.1364 0.1225 0.1603
0.15 0.1285 0.1072 0.1206 0.1247
0.0889 0.0983 0.0668 0.0499 0.0693
0.1
0.05
0
DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0

GRC
HUN IRL ITA LVA
LTU
0.6 0.5629 0.5871 0.5534
0.4796 0.5585
0.5 0.4846

0.4
0.3112
0.3 0.2907
0.2116 0.2365 0.2383 0.2254
0.2142
0.2 0.1295 0.1690 0.1587
0.0969 0.1060
0.1 0.0699 0.0535 0.0873 0.0530 0.0588
0.0449
0
DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0

POL PRT ROM SVK

MLT NLD
0.4565 0.4022 0.4246
0.6 0.5390 0.3852
0.3564 0.3371
0.5 0.3568 0.3270
0.3626
0.4
0.3 0.1530 0.2445
0.1279 0.1288 0.1723
0.2 0.0944 0.0885 0.1043
0.0705 0.0931
0.0639 0.0344 0.0525
0.1
0
DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0 DI<0 DI>=0

SVN SWE
0.6 0.5455

0.4
0.2520
0.2090
0.2 0.1379
0.0646 0.0949 0.0596

0
DI<0 DI>=0 DI<0 DI>=0

Source: WIFO calculations; BACI dataset (Gaulier and Zignago 2010)

59
are not used in the most efficient way. However, even
more importantly, the dynamic impact may affect the
adoption of new technology and further development
of capabilities. McMillan and Rodrick (2011) provide
evidence that countries with more flexible labour
markets experience growth-enhancing structural
change. Many factors can be identified, such as
certain types of taxes, labour market regulation, size-
dependent policies or trade barriers, in addition to
regulations and myriad costs of doing business in the
formal sector. Bartelsman et al. (2013) provide an
overall analysis that compares the United States to
seven European economies for the period 1992 to
2001 and find that idiosyncratic distortions play an
important role in the allocation of resources across
establishments. Their results suggest that output could
be increased by up to 15% in some countries if the
allocation of resources was improved. However, it is
very difficult to identify the sources of the
misallocation. One of the biggest impediments to the
reallocation of resources is financial frictions.
Financial markets are an important selection
mechanism for entrepreneurial projects and a well-
developed financial system is therefore important to
fostering entrepreneurial activity, structural change
and economic growth (Aghion et al. 2007, Buera et
al. 2011). Microeconomic evidence suggests that
credit market imperfections are important sources of
differences in productivity across countries. An
inefficient financial sector can significantly impede
the creation of new businesses and the growth of
enterprises. In particular, sectors with a larger scale
(e.g. manufacturing) and industries that have high
costs of product development (e.g. biotechnology) are
disproportionally affected by financial frictions.
However, financial repression that directs finance
towards certain sectors is not a force which supports
growth-enhancing structural change (Johansson and
Wang 2011). Institutional aspects such as government
effectiveness, low corruption and the efficiency of the
legal system are important to competitiveness in
terms of foreign direct investment (Alfaro et al.
2008). Thus, institutional quality is likely to affect
specialisation patterns.
Here the capabilities that affect specialisation and
structural change, which are associated with the
knowledge base of countries, are of greater
importance. The national innovation system
perspective also provides a useful view on these
issues as systemic failures are significant in
explaining the innovative performance of firms and
countries. The national system of innovation is
defined as a network of institutions in the public and
private sectors whose activities and interactions
initiate, import, modify and diffuse new technologies
(Freeman 1987). Systemic failures such as the lack of
interaction between the actors in the innovation
system, mismatches between basic research in
universities and applied research in industry,
malfunctioning technology transfer institutions, and
60
deficiencies in the absorptive capacity of enterprises
may all contribute to poor innovation performance.
Powell and Grodal (2005) show that innovation
networks have a positive impact on innovation
activity, but network failures can cause barriers to
innovation. Other evidence suggests that differences
in the patterns of technology diffusion may account
for a sizable part of the divergence in incomes
between rich and poor countries (e.g. Comin and
Mestieri Ferrer 2013).
Differences in the time scales of the adoption of new
technologies and the penetrations rates once new
technologies are adopted are important in determining
differences in economic structure. Here the lesson of
the literature on systems of innovation clearly
indicates that successful technology support policy
must consider arguments of systemic and institutional
failures. Growth traps and catch-up failures are most
often related to failures to select the right set of
institutions. For example, Acemoglu et al. (2006)
emphasise the different need for policy institutions
(educational systems, firm dynamics, innovation
policies) in countries that are close to or far from the
world technological frontier. Catching-up does not
depend on a particular institutional configuration, but
on the interlocking complementarities within the
institutional arrangements of the national innovation
system, an aspect that von Tunzelmann (2004) calls
network alignment. Structural change is thus
dependent on growth-enhancing policies and
institutions that allow the efficient allocation of
resources within economies. Policies and institutions
that hinder such reallocation processes are a prime
source of inefficiency and economic backwardness.
The recent crisis in Europe has shown that short-term
cyclical developments can lead to mispricing of assets
and a misallocation of economic resources, for
example with respect to the expansion of the
construction sector in the lead up to current crisis.
However, the evidence shows that the changes in the
manufacturing share are mainly related to the broad
trends of structural change mediated by international
specialisation documented earlier. For example, there
is nothing in the analysis of the inequality of
employment and value added shares, in Table 2.4 and
Table 2.5 that suggests that the divergence between
European countries increased substantially during the
crisis period (the drop in 2009 was symmetric across
countries and the manufacturing share normalised for
most countries in 2010 and 2011). Nevertheless, it is
known that some sectors such as manufacturing and
construction are very responsive to demand
downturns. The production of capital goods and
consumer durables are central industries in
manufacturing that are sensitive to changes in the
economic climate. Investment falls during recessions
as does business R&D (e.g. Aghion and Banerjee
2005, Hlzl et al. 2011). Thus fiscal policy measures
which aim at demand management should also
consider the structure of the economy and the pro-
cyclical behaviour of business R&D and innovation
activities over the business cycle. Supporting business
R&D and innovation during times of economic crisis
can support the ability of countries to achieve
economic growth in the long run and support
economic restructuring. Policies that aim at reducing
the openness of countries to international trade, in
contrast, are likely to be counterproductive. The
experience with financial repression and protection
from international competition is more often negative
than positive.
2.6. SUMMARY AND POLICY IMPLICATIONS
There are clearly identifiable broad patterns of
structural change that are quite homogenous across
countries and associated with the level of economic
development. The agriculture share is declining with
the process of economic development. The
manufacturing share is declining, both in terms of its
employment share and its value added share at a
certain point of economic development, while the
shares of services in employment and output are
increasing over time. These patterns can be broadly
explained in terms of the impact of technical change
and productivity improvements together with
changing patterns of demand due to higher income
(the Engels curve).
New technologies and new skills change
productivities, while demand patterns change with
changing income. The decline in the economic weight
of agriculture is associated with the increasing
mechanisation of agriculture which still leads to a rise
in labour and total factor productivity even in the
richest countries of the world. The increase in
productivity leads to lower prices and to a lower
factor demand (e.g. labour) if productivity
developments outstrip the growth in demand for
sectoral products. The same mechanism characterises
manufactured output, with the important difference
that the products (and the product characteristics) in
manufacturing are changing much faster than
products in agriculture. A hundred years ago there
was no computer industry and the output of the
electronics industry was very different.
However, if we look at shorter time periods, such as
the 15 years considered in the sectoral growth
decompositions, it can be seen that most growth
comes from processes of economic growth within
industries. It is a well-known stylised fact that
aggregate productivity improvements are mostly
related to within-sector (and even within-firm)
productivity improvements. Reallocation between
sectors and between industries becomes more
important the longer the time period of the analysis.
These patterns are almost the inevitable outcome of
the basic mechanisms underlying structural change.
61
Similar patterns are not only found using historical
data for European countries or cross-sectional data for
a large number of countries; these patterns are also
very similar for US states and EU members.
Nevertheless, it is equally important to realise that
there is some heterogeneity in the structure of
economies across countries. The working of structural
change is also mediated through international trade,
institutions, and international and domestic
competition.
On-going reallocations in economic weight across
different sectors should not be assessed only in terms
of reallocation of sector shares towards more
productive sectors. That would completely neglect the
linkages between sectors that are essential in
generating productivity improvement. Structural
shifts towards education-intensive activities (business
services and especially non-market services) do not
necessarily impact on growth potential in a negative
way, even if they apparently reduce aggregate
productivity. The education sector generates skilled
labour inputs for manufacturing and R&D.
International competitiveness is inter alia about trade
balances at the aggregate level; but creative
destruction at the product level is likely to be a
major driver of developments at the aggregate level.
Thus, international trade is an important determinant
of the development of sectoral shares in countries.
The successful catch-up stories of Germany in 19th
century and Japan and South Korea in the 20th
century cannot be explained without taking into
account international trade, comparative advantage in
tradables and specific competencies and capabilities
in the production of new and high-value added
products. Here it is important to acknowledge that
structural change shaping the economic development
of countries is highly path-dependent and cumulative.
Any change is rooted in present knowledge bases and
constrained by existing specialisation patterns.
Complementary capabilities need to be built up.
Therefore policies to support structural change should
always start by taking into account the existing
production structures of countries and regions, as well
as the knowledge base of supporting institutions.
Appropriate policies to foster structural change may
therefore also be country-specific and region-specific,
and depend on existing specialisation patterns. Skills
and technology are essential for achieving growth-
enhancing structural change. Structural change is
generally associated with the emergence of new
products and industries and the disappearance of
other products and occupations at the micro-
economic level which have a macroeconomic impact.
Producing more complex product classes and
upgrading existing products requires technological
competencies, skilled labour and administrative
capabilities at the business and government levels. It
should therefore not come as a surprise that the share
of services (non-government services such as
education and business services) starts to rise once
countries achieve income levels where the nature of
international competition changes from a purely cost-
driven to a more resource-intensive quality
competition. For the most successful exporting
countries it is crucial to develop new products that are
not produced by many other countries. Upgrading
possibilities are not distributed evenly: they seem to
be concentrated in high technology sectors and
complex products. Given the path-dependent
development of economic structures and comparative
advantage (as indicated by the product space
literature), countries seeking to shift their industrial
production up the technology ladder are likely to also
62
need to increase and improve non-government
services, such as education and business services.
The fact that upgrading structures is a cumulative
process makes it difficult to develop new
specialisation patterns out of the blue. This presents a
problem for countries where industrial restructuring is
necessary. The centrality of institutions and policies
in the process of structural change leads to a view that
the general quality of institutions is important to
structural change. Policies that foster structural
adjustments should therefore be conceived in a broad
way and cover such different areas as education,
research, technology and innovation policies, while
also focusing on the general quality of governance.
 REFERENCES
Acemoglu, D., Aghion, P. and Zilibotti, F. (2006), Distance to Frontier, Selection, and Economic Growth,
Journal of the European Economic Association, 4(1), pp. 3774.
Acemoglu, D., Johnson, S., and Robinson, J.A., (2001), The colonial origins of comparative development: An
empirical investigation, American Economic Review, 91(5), pp. 13691401.
Aghion, P., Fally, T., and Scarpetta, S. (2007), Credit Constraints as a Barrier to the Entry and Post-Entry
Growth of Firms, Economic Policy, 22, pp. 731779.
Aghion, P. and Banerjee, A. (2005), Volatility and Growth. Oxford: Oxford University Press.
Alfaro, L., Kalemli-Ozcan, S., and Volosovych, V., (2008), Why doesnt capital ow from rich to poor
countries? An empirical investigation, Review of Economics and Statistics, 90(2), pp. 347368.
Bartelsman, E., Haltiwanger, J., and Scarpetta, S., (2013), Cross-Country Differences in Productivity: The Role
of Allocation and Selection, American Economic Review, 103(1), pp. 305334.
Baumol, W.J., Blackman, S.A.B., and Wolff, E.N., (1989), Productivity and American Leadership: The long
view. Cambridge, MA: MIT Press.
Baumol, W.J., (1967), Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis, American
Economic Review, 57(3), pp. 415426.
Bolt, J., and Luiten van Zanden, J., (2013), The First Update of the Maddison Project; Re-Estimating Growth
Before 1820, Maddison Project Working Paper 4.
Bridgman, B., Dugan, A., Lal, M., Osborne, M., and Villones, S., (2012), Accounting for Household Production
in the National Accounts, 19652010, Survey of Current Business, 5, pp. 2336.
Buera, F.J., and Kaboski, J.P., (2012a), Scale and the origins of structural change, Journal of Economic
Theory, 147(2), pp. 684712.
Buera, F.J., and Kaboski, J.P., (2012b), The Rise of the Service Economy, American Economic Review, 102(6),
pp. 25402569.
Buera, F.J., Kaboski J., and Shin, Y., (2011), Finance and Development: A Tale of Two Sectors, American
Economic Review, 101(5), 19642002.
Clark, C., (1957), The Conditions of Economic Progress, 3rd edition, London: MacMillan.
Comin, D.A., Mestieri Ferrer, M., (2013), If Technology Has Arrived Everywhere, Why has Income
Diverged?, NBER Working Paper No. 19010.
Connolly, M., and Yi, K., (2008), How much of South Koreas growth miracle can be explained by trade
policy?, Working Paper Series 200823, Federal Reserve Bank of San Francisco
Cummins, J.G. and Violante, G.L. (2002) Investment-specific technical change in the United States (1947-
2000): Measurement and macroeconomic consequences, Review of Economic Dynamics, 5, 243-284.
Curtis, D.C.A., and Murthy, K.S.R., (1998), Economic growth and restructuring: a test of unbalanced growth
models, 19771992, Applied Economic Letters, 5(12), pp. 777780.
Dietrich, A., Krger, J.J., (2010), Long-run sectoral development: Time series evidence for the German
economy, Structural Change and Economic Dynamics, 21(2), pp. 111122.
Diewert, E., Fixler D., and Zieschang, K., (2012), Problems with the Measurement of Banking Services in a
National Accounting Framework, UNSW Australian School of Business Research Paper No. 201225.
Diewert, E., (2011), Measuring Productivity in the Public Sector: Some Conceptual Problems, Journal of
Productivity Analysis, 36, 177191.
Dollar, D., Kraay, A., (2001), Growth Is Good for the Poor, World Bank Policy Research Working Paper No.
2587.
Echevarria, C., (1997), Changes in sectoral composition associated with economic growth, International
Economic Review, 38(2), pp. 431452.
Eichengreen, B., and Gupta, P., (2013), The two waves of service-sector growth, Oxford Economic Papers,
65(1), pp. 96123.

63
Felipe, J., Kumar, U., Abdon, A., and Bacate, M., (2012), Product complexity and economic development,
Structural Change and Economic Dynamics, 23(1), pp. 3668.
Fourasti, J., (1949), Le Grand Espoir du XXe Sicle. Paris: Presses Universitaires de France.
Freeman, R.B., and Schettkat, R., (2005), Marketization of household production and the EU-US gap in work,
Economic Policy, 20(41), pp. 550.
Freeman, C., (1987), Technology Policy and Economic Performance: Lessons from Japan. London: Frances
Pinter.
Gaulier, G., and Zignago, S., (2010), BACI: International Trade Database at the Product-level. The 19942007
version, CEPII Working Paper No. 201023.
Gregory, M., and Russo, G., (2004), The Employment Impact of Differences in Demand and Production
Structures, DEMPATEM Working Paper No. 10.
Griliches, Z., (1992), Introduction, in: Zvi Griliches (ed.), Output Measurement in the Service Sectors. Studies
in Income and Wealth 56. University of Chicago Press for NBER, pp. 122.
Guerrieri, P. And Meliciani, V., (2005), Technology and international competitiveness: the interdependence
between manufacturing and producer services, Structural Change and Economic Dynamics, 16(4), pp. 489
502.
Herrendorf, B., and Rogerson, R., and Valentinyi, A., (2013), Growth and Structural Transformation, CEPR
Discussion Paper No. 9370.
Hidalgo, C. A., and Hausmann, R., (2009), The building blocks of economic complexity, Proceedings of the
National Academy of Sciences of the United States, 106(26), pp. 1057010575.
Hidalgo, C. A., Klinger, B., Barabsi, A.-L., Hausmann, R., (2007), The product space conditions the
development of nations, Science, 317, pp. 482487.
Hill, T.P., (1977), On goods and services, Review of Income and Wealth, 23(4), 315338.
Hlzl, W., Kaniovski, S., and Reinstaller, A., (2011), The exposure of knowledge and technology intense
sectors to the business cycle and their contribution to short run volatility and long run growth, in Janger, J.
(ed.), Structural Change and the competitiveness of EU Member States. Wien: WIFO.
Hlzl, W., and Reinstaller, A., (2007), The impact of productivity and demand shocks on structural dynamics:
Evidence from Austrian manufacturing, Structural Change and Economic Dynamics, 18(2), pp. 145166.
Isaksson, A., (2009), Structural Change and Productivity Growth: A Review with Implications for Developing
Countries, UNIDO Research and Statistics Branch Working Paper 08/2009.
Iscan, T.B., (2010), How Much Can Engel's Law and Baumol's Disease Explain the Rise of Service
Employment in the United States?, The B.E. Journal of Macroeconomics, 10(1), pp. 143.
Jorgenson, D.W., and Timmer, M.P., (2011), Structural Change in Advanced Nations: A New Set of Stylised
Facts, Scandinavian Journal of Economics, 113(1), pp. 129.
Johansson, A.C. and Wang, X., (2011), Financial Repression and Structural Imbalances, China Research
Centre, Stockholm School of Economics, Working Paper 201119.
Kaldor, N., (1996), Causes of Growth and Stagnation in the World Economy. Cambridge: Cambridge
University Press.
Kaldor, N., (1981), The role of increasing returns, technical progress and cumulative causation in the theory of
international trade and economic growth, Economie Appli u e, 34, pp. 593617.
Klimek, P., Hausmann, R., and Thurner, S., (2012), Empirical confirmation of creative destruction from world
trade data, PLoS ONE, 2012, 7(6), pp. 19.
Knack, S., and Keefer, P., (1995), Institutions and Economic Performance: Cross-Country Tests Using
Alternative Institutional Measures, Economics and Politics, 3.7, pp. 207228.
Kongsamut, P., Rebelo, S., and Xie, D., (2001), Beyond balanced growth, Review of Economic Studies,
68(237), pp. 869882.
Kuznets, S., (1966), Modern Economic Growth: Rate, Structure and Spread, New Haven: Yale University
Press.

64
Langbein, L., and Knack, S., (2010), The worldwide governance indicators: six, one or none? Journal of
Development Studies, 46(2), pp. 350370.
Lucas, R.E., (1988), On the mechanics of economic development, Journal of Monetary Economics, 22(1), pp.
342.
Maddison, A. (2010), Statistics on World Population, GDP and Per Capita GDP, 1-2008 AD, Groningen:
University of Groningen, 2010.
Matsuyama , K., (2009), Structural Change in an Interdependent World: A Global View of Manufacturing
Decline, Journal of the European Economic Association, 7(2-3), pp. 478486.
McMillan, M.S., and Rodrik, D., (2011), Globalization, Structural Change and Productivity Growth, NBER
Working Papers 17143, National Bureau of Economic Research.
Montresor S. and Vittucci Marzetti G. (2007), Outsourcing and structural change. What can input-output
analysis say about it?, Economia Politica, No. 1, pp. 43 78.
Ngai, R.L., and Pissarides, C.A., (2007), Structural change in a multisector model of growth, American
Economic Review, 97, pp. 429443.
Nordhaus, W.D., (2008), Baumols diseases: a macroeconomic perspective, The B.E. Journal of
Macroeconomics, 8(1), pp. 137.
Pasinetti, L., (1981), Structural Change and Economic Growth, Cambridge: Cambridge University Press.
Pasinetti, L., (1993), Structural Economic Dynamics: A Theory of the Economic Consequences of Human
Learning, Cambridge: Cambridge University Press.
Peneder, M., (2003), Industrial Structure and Aggregate Growth, Structural Change and Economic Dynamics,
14(4), pp. 427448.
Peneder, M., Kaniovski S., Dachs, B., (2003), What Follows Tertiarisation? Structural Change and the Role of
Knowledge Based Services, The Service Industries Journal, 23(2), pp. 4766.
Powell, W.W., and Grodal, S., (2005), Networks of Innovators, in Fagerberg et al. (eds.) The Oxford Handbook
of Innovation. Oxford: Oxford University Press.
Pugno, M., (2006), The service paradox and endogenous economic growth, Structural Change and Economic
Dynamics, 17(1), pp. 99115.
Reinstaller, A., Hlzl, W., Kutsam, J., and Schmid, C., (2012), The development of productive structures of EU
Member States and their international competitiveness, Report prepared under Specific Contract No SI2-
614099 implementing the Framework Contract ENTR/2009/033 for the European Commission, DG
Enterprise and Industry.
Nordhaus, W.D., (2008), Baumols diseases: a macroeconomic perspective, The B.E. Journal of
Macroeconomics, 8(1), pp. 137.
Pasinetti, L., (1981), Structural Change and Economic Growth, Cambridge: Cambridge University Press.
Pasinetti, L., (1993), Structural Economic Dynamics: A Theory of the Economic Consequences of Human
Learning, Cambridge: Cambridge University Press.
Peneder, M., (2003), Industrial Structure and Aggregate Growth, Structural Change and Economic Dynamics,
14(4), pp. 427448.
Peneder, M., Kaniovski S., Dachs, B., (2003), What Follows Tertiarisation? Structural Change and the Role of
Knowledge Based Services, The Service Industries Journal, 23(2), pp. 4766.
Powell, W.W., and Grodal, S., (2005), Networks of Innovators, in Fagerberg et al. (eds.) The Oxford Handbook
of Innovation. Oxford: Oxford University Press.
Pugno, M., (2006), The service paradox and endogenous economic growth, Structural Change and Economic
Dynamics, 17(1), pp. 99115.
Reinstaller, A., Hlzl, W., Kutsam, J., and Schmid, C., (2012), The development of productive structures of EU
Member States and their international competitiveness, Report prepared under Specific Contract No SI2-
614099 implementing the Framework Contract ENTR/2009/033 for the European Commission, DG
Enterprise and Industry.

65
Restuccia D., and Rogerson, R., (2013), Misallocation and productivity, Review of Economic Dynamics, 16(1),
pp. 110.
Reyes-Heroles, R., (2012), Structural Change in an Open Economy: A Quantitative Exploration, Princeton
University.
Rogerson, R., (2008), Structural transformation and the deterioration of European labor market outcomes,
Journal of Political Economy, 116(2), pp. 235259.
Rowthorn, R., and Ramaswamy, R., (1999), Growth, Trade and Deindustrialization, IMF Staff Papers, 46 (1),
pp. 1841.
Schettkat, R., and Yocarini, L., (2006), The shift to services employment: A review of the literature, Structural
Change and Economic Dynamics, 17(2), pp. 127147.
Triplett, J.E., and Bosworth B.P., (2004), Productivity in the U.S. Services Sector: New Sources of Economic
Growth. Washington D.C.: Brookings Institution Press.
Uy, T., Yi, K-M., and Zhang, J., (2012), Structural Change in an Open Economy, Manuscript University of
Michigan.
von Tunzelmann, N., (2004) Integrating Economic Policy and Technology Policy in the EU, Revue d'conomie
industrielle, 105(1), pp. 85104.
Yi, K-M., and Zhang, J., (2011), Structural Change in an Open Economy, Staff Report 456, Federal Reserve
Bank of Minneapolis.

66
 ANNEX 1
DEFINITION OF PRODUCT SPACE INDICATORS
Product space indicators
Hidalgo et al. (2007) define the product space as a bi-partite network linking countries to products. To construct
this network, they define a proximity measure, , between two products i and j as the pairwise conditional
probability P of a country exporting one good given that it exports another. This measure is defined as follows:
{ ( | ) ( | )} (proximity)
where RCAi means that a country has a revealed comparative advantage (RCA) for product i and is therefore a
significant exporter of that product. The RCA is taken in order to ensure that marginal exports do not introduce
noise into the data. The minimum is taken to avoid that if a country would be a sole exporter of a good the
conditional probability would take on the value 1. By taking the minimum of the reciprocal relationship this
problem is avoided. Proximity is therefore a measure that links any product to any other product traded in the
world. In terms of a network, the proximity can be conceived as the edges of the network with the products being
its nodes.
In order to assess the likelihood that a product becomes a significant export in a country Hidalgo et al. (2007)
define a measure called density. We refer to this indicator as neighbourhood density to distinguish it from
the statistical notion of density. It measures the average proximity of a product to a countrys current productive
structure. For products for which the country is not a yet a significant producer this measure therefore indicates
how embedded the product would be and by implication to what extent complementary capabilities are already
available in a country. It therefore captures the likelihood that a country develops a comparative advantage in
any product. The neighbourhood density is calculated as follows:
(neighbourhood density)
where is unity if product i has an RCA>1 in country k. The neighbourhood density takes on the value 1 if a
country produces all i products to which product j is connected in the product space. The neighbourhood density
is therefore normalised between 0 and 1 and takes on the maximum when a product is connected to all other
products in the product mix of a country.
The product complexity scores (PCS) have been calculated using the method of reflections advanced by Hidalgo
- Hausmann (2009). It exploits information on the diversification of a country and the ubiquity of the products
(i.e. in how countries have an RCA>1 for a given product). is the matrix linking product to countries and
has an entry of 1 if country c has an RCA for product p. Then the matrix can be summed up row wise over
products p one obtains a measure for the diversification of a country c.

If, on the other hand, the matrix is summed up column wise one obtains a measure for the ubiquity of
comparative advantage in the trade of a specific product p, i.e. this measure tells us how many countries c have a
comparative advantage in trading this product.

By combining these two indicators it is possible to calculate through recursive substitution how common
products are that are exported by a specific country,

and how diversified the countries are that produce a specific product

If formula (3) goes through an additional iteration the indicator now tells us how diversified countries are that
export similar products as those exported by country c. An additional iteration for formula (4) tells us then how
ubiquitous products are that are exported by product ps exporters. Each additional iteration n adds information
on the neighbour of a country or product that is n steps away from country c or product p. Higher iterations than

67
those presented in the table are increasingly difficult to interpret. The indicator standardised relative to all
products at iteration n gives the product complexity score PCS.
It is possible to calculate the simultaneous or slightly lagged appearance and disappearance of products using
product space metrics. In this way a dynamic view is introduced in the product space analysis. We follow the
method proposed by Klimek et al. (2012). To calculate the co-appearance and displacement indices define
appearance and disappearance events as follows:
Appearance: .
Disappearance: ,
where t defines a specific point in time and c a specific country, is the value of exports a country has in
any product class i. The product index i,j runs from 1 to n, where n corresponds to the number of product classes
in the analysis. Hence, the empirical number of co-appearances between any pair of product classes i and j is
given by

whereas the empirical number of displacements over period after the appearance of a product class i is given
by

The period was set to 3 such that all displacements of a product class j three years after the appearance of
product class i have been taken into account.
The co-appearance index AI follows then from equation (1) if on the one hand we control for the fact that
products with a high number of appearance are likely to have also a higher number of co-appearances, and if on
the other hand the resulting factor is normalised to lie in the interval[ ].

[ ]
where is the number of appearances of each product class i and j across countries c and
over all observation periods t, and is the normalisation factor rescaling the sum to the established range.
The displacement index DI is instead defined as

[ ]

Clearly, if the sum in equation (4) is negative, then product class i is on average displaced more often by
appearances of the other product classes j during the period . A positive indicator value instead means that i
displaces on average more often any other product class j than j replaces i after its appearance. is again the
normalisation factor rescaling the sum to the established range.
In order to analyse the displacement across NACE sectors we have aggregated the displacement scores
as follows:

where weights represent the share in total export value of sectors k and l of products i and j in Sector S. The
weights therefore give higher importance to product class displacements that have a higher value in total exports
of a NACE sector than those that have a lower value. For all calculations period has been set to 3, i.e. we
include all disappearance events three years after the appearance of a product in the counts.

68
 Chapter 3.
REDUCING PRODUCTIVITY AND EFFICIENCY GAPS:
THE ROLE OF KNOWLEDGE ASSETS, ABSORPTIVE
From the mid-1990s productivity growth in the EU
slowed down compared to the US, which in contrast
was experiencing rapid productivity acceleration
(OMahony and van Ark, 2003). As a result, the US-
EU productivity gap widened. While the post-1995
productivity slowdown was felt across all EU
countries, productivity trends have differed and
economic disparities have amplified since the
economic and financial crisis, which hit countries
with different intensities (Mas, 2010). Understanding
the reasons underlying productivity differentials has
become a priority for policy makers so that useful
policies to promote and restore long-lasting economic
growth in Europe, beyond the traditional models of
catch-up and convergence, can be implemented.
Productivity, a key source of economic growth and
competitiveness, is defined as the amount of output
that can be produced per unit of input. The term
productivity, however, is used to describe two related
concepts: labour productivity and total factor
productivity (TFP). Labour productivity, which refers
to the amount produced per each unit of labour, can
be improved through either a greater use of capital
relative to labour (capital deepening), or through an
increase in TFP growth. TFP measures the part of an
output increase not accounted for by increases in the
quantity and quality of inputs. TFP movements are
mainly due to technical efficiency increases, which
imply catching up to the existing technology frontier,
or due to technological improvements as the frontier
shifts outwards over time.
Prior to the financial and economic crisis which
started during 2007-2008, the debate on the European
labour productivity slowdown pointed to ICT capital
accumulation as a major reason for the under-
performance of EU labour productivity. This largely
reflected the significantly slower adoption of ICT
technologies in the EU compared to the US, in
particular in services sectors. Moreover, industry-
based studies revealed that the US productivity
advantage was concentrated in specific services
sectors, mainly trade, finance and business services
(Timmer et al. 2010). In these ICT-intensive using
sectors, the large ICT investment flows during the
second half of the 1990s together with
complementary investments in organizational capital
led to a rapid TFP growth during the first half of the
2000s (Van Ark et al, 2008; Brynjolfsson and
Saunders 2010).
69
CAPACITY AND INSTITUTIONS
The initial hypothesis was that Europe was merely
lagging behind the US in the adoption of ICT
technologies, and therefore, it would take some time
for its benefits to materialise. Now, it has become
apparent that high levels of investment alone do not
produce faster economic growth and better
productivity performance. Several years after the
ICT revolution, the EU is not only still lagging
behind the US, but the productivity growth gap has
recently widened.
Empirical findings show that in the EU there was
insufficient investment in the skills and
organizational changes necessary to reap the benefits
of ICT technologies (Brynjolfsson and Hitt 2000;
OMahony and Vecchi 2005). Lower investments in
intangible assets broadly conceived (R&D, human
capital, etc.) are likely to explain a portion of the US-
EU productivity gap as these factors affect countries'
absorptive capacity, i.e. their ability to take advantage
of the technology developed elsewhere (international
technology transfers).
Most of the leading technologies available worldwide
are developed by a few frontier countries which
dominate the entire global market. Technological
laggards can benefit by imitating such technologies
through international trade. However, in order to
assimilate and exploit the foreign knowledge in the
production of their own goods, it is indispensable for
laggard countries to develop a certain degree of
absorptive capacity, i.e. to reach a minimum
threshold of technological competence. Absorptive
capacity is considered essential to close the gap with
the technology leaders and spur economic growth
(Griffith et al. 2004).
Another factor that has recently been identified as a
cause of the lower TFP performance in the EU, is the
more rigid regulatory framework compared to that in
the US (Nicodme and Sauner-Leroy 2004, Bourls
et al. 2012). For example, it has been shown that low
levels of competition and strict employment laws
prevent the necessary optimal adjustments to factor
allocation in order to take full advantage of new
technologies (Conway et al. 2006; Bassanini et al.
2009; Arnold et al. 2011).
An issue largely unexplored, and to which this study
wishes to contribute, is whether the regulatory
environment determines the efficiency with which
resources are used in production (technical
efficiency). This has become a major issue in recent
years as the ability to exploit existing resources
emerges as one of the most important sources of
productivity gains in the most mature economies (van
Ark et al., 2012).
Although the recent downturn has shifted the focus
on the functioning of markets as, a possible
mechanism driving productivity differentials across
areas of the world, understanding the channels
through which the regulatory environment determines
TFP growth, remains a challenge.
The main contribution of this chapter is to provide a
comprehensive analysis of the determinants of
productivity growth, focusing on the role played by
the restrictions in the product, labour and financial
markets as well as, the role of intangible assets (e.g.
ICT, R&D) and absorptive capacity (e.g. skills). The
use of Stochastic Frontier Analysis, is used to
investigate the main factors affecting changes in
technical efficiency for a large sample of countries
and industries in the EU, an issue largely unexplored
in the literature to date.
Specifically, this chapter addresses the following
questions:
- What are the recent trends in productivity growth
in the EU and the US? Which areas of the
economy are driving the most recent growth
patterns? What is the relative role of the various
factors inputs?
- Do the institutional framework and laws
governing the functioning of the factor and
product markets shape the EUs ability to benefit
from technology originated in the frontier
countries?
- Does the EU regulatory setting influence the
level of production efficiency, i.e. the ability of
firms/industries to use factor inputs in the most
technically feasible way? What are the main
institutional factors that explain the efficiency
gap with the US? To what extent do these factors
interact with the use of ICT and with the
technology characteristics of EU industries?
Section 3.1 of this chapter highlights the main
productivity and growth trends in the EU in
comparison with other major world economies for the
period from 1995 to 2012. A decomposition of labour
productivity growth into its main components, and a
70
detailed up-to-date account of sectoral productivity
developments are provided. Section 3.2 reports
econometric evidence on the factors affecting
international diffusion of R&D focusing on the
institutional determinants of a countrys absorptive
capacity. Section 3.3 quantifies the extent to which
ICT and institutional factors have had an impact on
the efficient use of resources. Assessing both the
economic and institutional drivers of technical
efficiency is helpful to understand the sources of the
productivity gap and to design policies that might
reduce it. Section 3.4 integrates the analysis by
presenting evidence on firm behaviour at the outset of
the crisis, focusing on firms strategic decisions
regarding investments in tangible and knowledge
assets and their impact on productivity. Section 3.5
concludes the analysis and outlines the policy
implications.
3.1. GROWTH ACCOUNTING AND THE EFFECT OF
THE CRISIS AT COUNTRY AND SECTOR LEVEL
3.1.1. Economic performance of the EU and
other major economies: overview of
aggregate output and productivity
trends
This section presents an overview of recent output
and productivity trends in the EU, highlighting the
main convergence and divergence patterns from a
comparative perspective. From the mid-1990s the US
economy has grown at a higher rate than the EU and
Japan (see Figure 3.1). During the period 1995-2004,
the average GDP58 growth rate in the US was 3.3%,
around 0.85 percentage points higher than that
experienced by the EU-27. This trend reversed
briefly during the period 2004-2007, as the EU-27
started to grow faster than the US (3% versus 2.5%).
This performance was partly driven by the newest
Member States, as GDP growth in the EU-27 was
higher than the EU-15. Japan performed considerably
worse than Europe during the period 1995-2007,
achieving only moderate GDP growth rates (between
1 % and 1.5%).
At the outset of the crisis, in 2008, output growth
slowed down across all areas, and in 2009 output
levels fell globally. By 2010, however growth had
resumed across the US, EU and Japan. The US
exhibited the strongest recovery. Performance in the
EU improved during 2010 and 2011; however, the
sovereign debt crisis caused a fall in GDP growth in
2012. Japans output level has remained largely flat
since 2010.
58
The source for the Gross Domestic Product data is the Total
Economy database, The Conference Board, January 2013
release.
Figure 3.1. GDP growth in the EU, US and Japan. 1990-2012 (1995=100)
160

150

140

130

120

110

100

90

80

70
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

EU-27 US JP EU-15

Source : The Conference Board Database and own calculations.

Figure 3.2. GDP per hour growth in the EU, US and Japan. 1990-2012 (1995=100)
140

130

120

110

100

90

80

70
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

EU-27 US JP

Source: The Conference Board Total Economy Database and own calculations.

71
Figure 3.2 illustrates trends in productivity, measured factors contributing to labour productivity growth are
as GDP per hour, in the EU, US and Japan. From the shown in Figure 3.3 for Europe and Figure 3.4 for the
mid-1990s to the early 2000s productivity accelerated US.
in the US but not in Europe or Japan. In a period
characterised by the widespread diffusion of Figure 3.3. Growth accounting in the EU: Decomposition
Information and Communication Technologies (ICT), of labour productivity growth, 1995-2012
the US productivity lead was amplified thanks to its
greater ability to invest and benefit from the new
technology. While EU productivity showed signs of
catching up towards US levels during 2004-2007 (see
Figure 3.1), the onset of the crisis has worsened the
EU position. Although slower than in the pre-crisis
period, the US labour US labour productivity growth
rate fluctuated around 1% per annum, even at the
height of the global downturn. In contrast, in the EU
and in Japan, labour productivity levels fell in the
aftermath of the crisis, only to recover from 2010. As
a result, the gap in labour productivity between the
Source: EUKLEMS and own calculations.
US and the EU widened again. Between 2007 and
2012, the US labour productivity growth was During the period 1995-2004, the most important
approximately 2 percentage points higher than contributor to labour productivity in the EU59 was the
Europe and 3 percentage points higher than Japan. accumulation of capital assets. TFP gains in the EU
From 2010 onwards Japan experienced a strong were significant during this period but growth rates
productivity recovery, outperforming the EU and the were considerably lower than in the US. In the US,
US. This contrasts with the slowdown in output the main factor driving labour productivity growth
growth shown in Figure 3.1. This finding is a during the period 1995-2004 was ICT capital. The
consequence of the large reduction in the total growth contribution of ICT was substantially slower
number of hours worked, which testifies the in the EU.
flexibility of the Japanese wage system and its labour
hoarding tradition (Darby et al. 2001). In the case of Figure 3.4. Growth accounting in the US: Decomposition
the EU, which is characterised by a large degree of of labour productivity growth, 1995-2012
heterogeneity in the institutional and policy
environment, labour market responses were largely
country-specific. While some countries were able to
adjust to worsening demand conditions via a
reduction in hours worked (e.g. UK, Germany,
France, Netherlands) others carried out major labour
shedding, mostly concentrated in low-skill sectors,
resulting into an overall increase in productivity
levels (e.g. Spain and Ireland).

3.1.2. Growth accounting analysis: Sources of

productivity growth at aggregate level Source: EUKLEMS and own calculations.

The EU experienced significant TFP acceleration

The objective of this section is to explore the role of
during the subsequent period, 2004-2007. Non-ICT
inputs to production and Total Factor Productivity
capital accumulation and TFP were the main factors
(TFP) in explaining aggregate labour productivity
contributing to labour productivity growth and to the
developments, measured by GDP per hour. The
catching up process towards the US, shown in the
inputs considered are capital assets, distinguished into
main output trends. During the same period, the US
ICT and non-ICT assets, and labour composition,
experienced a productivity slowdown, mainly caused
which represents the contribution of skilled labour.
by a decrease in the speed of ICT capital
The TFP component, which is derived using the
accumulation and a lower contribution of TFP
neoclassical growth accounting methodology,
growth.
quantifies the part of the output growth not accounted
for by growth in the quantity and quality of inputs;
TFP captures the influence of unmeasured factors on
59
productivity, such as efficiency improvements, The EU includes the following eight countries: AT, BE, DE,
ES, FR, IT, NL, UK.
technological change and spillovers. The different

72
After 2007, the effect of the crisis was particularly
strong in Europe and negatively affected the
contribution of all factors of production, with the sole
exception of labour composition. The latter finding is
consistent with prior evidence of labour quality
growth in many European countries. For example,
Kang et al. (2012) find that the average skill level of
the workforce, which mainly reflects qualifications
achieved through the general education system, rose
during the recession years. TFP growth was mostly
affected by the financial crisis as its contribution
turned negative. The contribution of capital, although
still positive, declined substantially compared to the
pre-crisis period, reflecting the consequences of
tightening credit conditions for European firms.
Goodridge et al. (2013) observed that in the UK,
while capital investment decreased considerably with
the recession, investments in intangible assets
increased, particularly investments in R&D and
software. This, together with a higher proportion of
skilled workers, increases future growth potential,
facilitating recovery.
In the US, the crisis also affected the contribution of
TFP and non-ICT capital, although neither turned
negative. The contribution of ICT capital decreased
slightly compared to the 2004-2007 period, while
labour composition increased slightly. This testifies
that, as in the UK, US firms prioritised the
employment of highly skilled workers.
The analysis of growth trends at the aggregate level is
consistent with the hypothesis that the US
productivity lead in the late 1990s and early 2000s
resulted from a first mover advantage in ICT. Large
TFP gains at the time of rapid ICT investment
suggests that ICT may have had an impact on
productivity beyond that of ICT-capital deepening
(OMahony and van Ark 2003); this has been
attributed, for instance, to the existence of spillovers
related to knowledge assets and to large investments
in organizational capital (Bryonjolffson and Saunders
2010). The EU experience reveals that TFP
movements may have followed the US productivity
developments, albeit with a time-lag. Basu et al.
(2004) argue that the `missing TFP growth in the
United Kingdom, compared to the US, in the second
half of the 1990 is likely to have been caused by a
delay in undertaking investments that would
complement the adoption of the new technology.
In the EU, substantial declines in TFP have been
recorded since 2007. TFP appears to have behaved in
a highly pro-cyclical way and existing contributions
suggest that this reflects a decline in the overall
efficiency of the production process. However, these
results should be treated cautiously as it is too soon to
73
draw conclusions based on the most recent TFP
movements (OECD, 2012).
3.1.3. Productivity developments in the EU and
the United States: a sectoral perspective
A more detailed explanation of the nature of the EU-
US productivity gap is provided by looking at the
contribution of each sector in the economy. Labour
productivity growth trends are examined for specific
industries (Figure 3.6. and Figure 3.5.). While the
analysis of these sectoral productivity trends are
informative, offering an interesting snapshot of
pockets of growth, the relative industry contribution
to the overall EU-US productivity gap will be
determined by the size of each sector and by the
differences in industrial structure.
Due to limited data availability60, the focus here is on
a group of eight EU countries61 (Austria, Belgium,
Spain, France, Germany, Italy, Netherlands and the
UK). Industry productivity data are drawn from the
latest release of the EUKLEMS database (OMahony
and Timmer 2009) and follow the NACE Rev. 2
classification of economic activities.
During the period 1995-2004, the sectors
experiencing the highest growth rates in the US
include the ICT-producing sector electrical and
optical equipment, with an average growth of almost
21%, followed by coke and refined petroleum
products62, with a rate of 15%. Then come
information and communication activities63 and
wholesale and retail activities, which experienced
productivity growth rates of around 4%.
In the EU, the best performing sectors during the
same period included coke and refined petroleum,
with a rate of labour productivity growth rate of 8%,
finance and insurance activities, and electricity, gas
and water supply, both with rates of around 6%. Next
were chemicals and chemical products, with a rate of
5%. During that period in the EU the lowest labour
productivity growth rates were observed in retail
trade, professional, scientific, technical,
administrative and support services; community,
social and personal services; and arts, entertainment,
recreation and other service activities.
60
EUKLEMS most recent updates, covering up to year 2010 are
only available for a limited number of European countries.
61
These economies represented in 2012 approximately the 80%
of EU output.
62
Care needs to be taken in interpreting these results as
measurement issues in this sector may be important.
63
The Information and Communication sector (J code in Nace
Rev. 2) comprises the following activities: publishing,
audiovisual and broadcasting activities; telecommunications;
IT and information services activities.
Between 1995 and 2004, the European failure to
match the US acceleration in output and productivity
has largely been attributed to developments in market
services (Timmer et al. 2010). The analysis in this
section reveals that the sector which most contributed
to amplify the US productivity advantage was in fact
wholesale and retail distribution, due to its strong
productivity performance and its relatively large
share in the economy. Other services sectors with
sizeable contributions include the professional,
scientific, technical, administrative and support
services, and finance and insurance activities. These
findings are consistent with previous evidence (EC
2008). The electrical and optical equipment sector
also made a key contribution with its outstanding
growth performance.
Throughout the period 2004-2007, two factors jointly
contributed to the reduction of the EU-US
productivity gap: the acceleration of productivity in
most EU manufacturing industries relative to the
productivity performance in US manufacturing, and a
robust performance of many EU services sectors.
The highest productivity growth rates in the EU
manufacturing sector64 were achieved in chemicals
and chemical products, with a rate of 13%; in
contrast, the US electrical and optical equipment
sector continued to show impressive growth while in
the EU growth remained modest.
In services, labour productivity slowed significantly
between 2004 and 2007 in the US wholesale and
retail sector compared to the exceptional performance
observed in the earlier period. On the other hand, the
EU performance in the same sector improved
substantially, reaching a 3% productivity growth,
nearly doubling the growth rate achieved in previous
periods. The information and communication
activities sector experienced robust labour
productivity growth in both the US and the EU; in the
latter area, this meant a considerable improvement as
labour productivity had previously followed a
deteriorating trend. In most EU services, labour
productivity improved, particularly in the
professional, scientific, technical activities, and
community, social and personal services. This is in
contrast to the poor performance of both sectors in the
period 1995-2004. Overall the evidence shows that
those sectors which contributed to narrowing the EU
productivity gap relative to the US between 2004 and
2007, were those responsible for the stagnant EU
productivity in the previous decade.
64
The coke and refined petroleum products sector (code 19 in
Nace Rev.1) is excluded from this picture.
74
During the financial crisis (2007-2010), labour
productivity stalled in the EU, while it continued to
improve in the US. The majority of manufacturing
sectors in the EU-8 experienced a fall in productivity
levels, probably reflecting a higher exposure to global
demand fluctuations compared to the services sectors.
Manufacturing productivity as a whole decreased by
more than 1% annually, with chemicals, decreasing
by more than 4% annually. Productivity in the
construction sector also deteriorated considerably as
well as in some services activities, such as wholesale
and information and communications. Those sectors
that showed the most resilience to the weakening
economic conditions in the EU-8 were financial and
insurance activities growing by 6% annually,
professional, scientific, technical activities and retail
trade growing by around 2% per annum.
In the US, manufacturing productivity grew by over
4% annually during 2007-2010. One of the few
sectors that experienced a worsening in productivity
levels was chemicals. The majority of services
activities though experienced robust growth, in
particular telecommunications, finance and insurance
and IT and information services. Productivity growth
in the electrical and optical equipment sector did not
show signs of slowing down.
Other interesting lessons can be drawn from the
analysis of post-crisis industry trends. In summary,
the sectors which contributed to further increase the
US productivity advantage are electrical and optical
equipment and the majority of manufacturing sectors,
as well as construction, and telecommunications
(which had shown an outstanding performance in the
EU prior to the crisis). Those sectors which helped
the EU to narrow the gap in the most recent period
include financial activities and business services,
accommodation and food, some public services and
other services activities. Many of the ICT-using
services sectors that had improved their labour
productivity in the pre-crisis years continued to
perform well; the exception was wholesale and retail,
greatly affected by weak consumer demand.
A further extension of the industry level analysis
based on growth accounting allows the identification
of those factors that played a major contribution in
determining productivity growth performance in
selected industries. The results, discussed in detail in
the background study, reveal that up to 2007 TFP was
the main driver of productivity growth in the EU and
US manufacturing. In the EU, TFP was also the main
source of declining productivity trends after the crisis,
next to physical capital contributions. In services the
picture is more heterogeneous. In wholesale and
retail, the contribution of TFP in Europe is delayed
compared to the US, and it particularly affects
productivity in the latest years before the financial
Figure 3.5. Sectoral labour productivity growth rates, US, 1995-2010
1995-2004 2004-2007 2007-2010
25

20

15
Percentage

10

0
A

28

61
D-E
B
10-12
13-15
16-18
20-21
22-23
24-25
26-27

29-30
31-33

F
G
H
I
58-60

62-63
K
L

R-S
M-N
O
P
Q

R
S
-5

-10

-15
Source: EUKLEMS and own calculations.

Figure 3.6. Sectoral labour productivity growth rates, EU-8, 1995-2010

1995-2004 2004-2007 2007-2010

25

20

15

10
percentage

0
10-12
13-15
16-18
20-21
22-23
24-25
26-27

58-60
28
29-30
31-33

61
62-63
D-E
A
B

F
G
H
I

K
L

O
P
Q
R-S
R
S
T
M-N

-5

-10

-15
Source: EUKLEMS and own calculations.

crisis. In information and communication services, productivity performance since the crisis, this sector
the contribution of TFP was particularly large since continues to fare considerably well.
the mid-1990s in both countries. Despite a declining

75
3.2. THE ROLE OF KNOWLEDGE TRANSFER,
ABSORPTIVE CAPACITY AND INSTITUTIONS FOR
PRODUCTIVITY GROWTH
One way to better understand possible causes of the
productivity gap is to consider the impact of
investments in intangible assets such as R&D. The
importance of investing in innovation activity has
long been recognised in the theoretical and empirical
literature (Romer 1990, Aghion and Howitt 1992,
Park 2008, Nishioka and Ripoll 2012). What is
perhaps less often acknowledged is that resources for
such innovations tend to be highly concentrated in a
small number of advanced OECD countries65, which
have the required skills and institutions in place to
invest heavily in R&D. This implies that for countries
whose firms are not at the technological frontier, the
diffusion of technology from the frontier is likely to
be an important source of productivity growth,
through both imitation and follow-on innovation and
adaptation (Evenson and Westphal, 1995).
Knowledge transfers can occur via different channels,
such as FDI, joint ventures, reverse engineering, and
collaborations. A survey of the biggest EU R&D
investing companies shows that knowledge transfer is
more important among companies than between
companies and the public sector. Knowledge transfer
is especially relevant for companies in high R&D
intensity sectors66. In this section the focus is on the
diffusion of technology via intermediate goods trade
an approach consistent with existing theoretical
models (e.g. Grossman and Helpman 1991). The first
objective of the analysis is to assess how knowledge
transfers affect productivity growth; secondly, the
role of absorptive capacities and barriers to diffusion
is taken into account, as these can make an important
difference to the extent to which countries benefit
from innovations carried out elsewhere.
The methodology used to study the impact of foreign
R&D spillovers follows closely the contribution by
Nishioka and Ripoll (2012), which requires data on
R&D stocks by country and industry and input-output
tables capturing inter-industry and inter-country
linkages. This is carried out using data on R&D
65
The share of R&D financed by enterprises in advanced
countries was 98% in the 1980s and 94% in the 1990s
(UNIDO, 2002). Even within developed countries however
R&D is concentrated, with Eaton and Kortum (1999) noting
that in the late 1980s, 80 percent of OECD research scientists
and engineers were employed in five countries (US, UK,
Germany, Japan and France).
66
See: www.jrc.es: Tbke, A.; Hervs, F. and Zimmermann, J.:
"The 2012 EU Survey on R&D Investment Business Trends",
European Commission, Joint Research Centre, EUR 25424
EN, pp.21.
76
expenditure from the OECD ANBERD database67,
from which a R&D stock for ten manufacturing
industries and 20 countries is calculated using the
perpetual inventory method. Information on
intermediate flows required for the calculation of the
R&D stock of intermediates is taken from the
recently compiled World-Input-Output-Database
(WIOD), which reports data on socio-economic
accounts, international input-output tables and
bilateral trade data across 35 industries and 41
countries over the period 1995-2009 (see
Dietzenbacher et al. 2013)68.
Though R&D stocks could only be calculated for a
limited number of countries and industries, two
interesting stylised facts emerge: firstly, over 80% of
the R&D stock is concentrated in a small number of
industries (electrical and optical equipment, transport
equipment and chemicals and chemical products and
in particular the pharmaceutical sector which plays a
key role among highly innovative industries);
secondly, the US and Japan dominate R&D stocks in
the sample of countries considered, with respectively
40% and 28% of the total, followed by Germany
(11%), France (8%) and the UK (7%). This indicates
that these five countries account for 80% of the
overall R&D stock, consistent with Eaton and
Kortum (1999).
Using these R&D stocks and the information on inter-
industry and inter-country linkages makes it possible
to calculate the variables capturing the direct and
indirect R&D content of intermediate input flows:
First, the R&D stocks divided by gross output
provides a vector containing the direct R&D
requirements by sector and country. This vector is
then multiplied with the global Leontief inverse,
derived from the WIOD and the global inter-industry,
transaction matrix. The latter only includes the
foreign inter-industry flows therefore capturing the
role of international R&D spillovers.69
As expected, those countries and industries which
dominate R&D expenditures also tend to have the
highest shares in the direct and indirect use of R&D70.
67
It is important to stress that the BERD data are territory based.
Alternatively, attention might be paid to this issue at the
company level (e.g. Cincera and Veugelers, 2010).
68
Some of the associated data have been updated to 2011.
69
Technical details are provided in the background study to this
chapter.
70
It should be stressed here that these results are confined to
those countries and industries for which reliable R&D data are
available. As R&D is concentrated in a few industries and
countries, the computation of direct and indirect R&D stocks
provide good proxies for international R&D spillovers that can
be used in empirical econometric analysis. However, no more
detailed inference on source and use country and industries can
be made without having better knowledge of R&D stocks
across all countries and industries.
Relatively large shares are also found for
construction, suggesting that there are strong linkages
between manufacturing and this sector. With respect
to countries, large increases in the indirect use of
R&D can be found for the US, Japan and the rest of
the world between 1995 and 2010. China has also
experienced a large increase in both direct and
indirect R&D usage over time. This increase is solely
due to increased flows of R&D intensive
intermediates into China over this period (China is
excluded from the list of R&D source countries given
the limited industry data).
The assessment of the role of absorptive capacity and
barriers to diffusion requires the inclusion of
additional variables. Absorptive capacity is measured
using information from the 2013 release of Barro and
Lees dataset on the average years of secondary
schooling in the population over 15 years of age
(Syr)71. As additional indicator of absorptive capacity,
the analysis includes the level of R&D taken in logs
(Cohen and Levinthal 1989). Data on R&D are
extracted from the OECD ANBERD dataset. The role
of absorption barriers is assessed by using several
OECD indicators. A first set describes regulation in
the labour market and includes: an indicator on
strictness of regulation of employees on regular
contracts (EPR), an indicator for strictness of
regulations for temporary forms of employment
(EPT), and an indicator of strictness of regulation and
specific requirements for collective dismissal (EPC).
These variables are on a scale of 0 to 6, with 0 having
the least and 6 the most restrictions. To be consistent
with the hypotheses of Parente and Prescott (1994,
1999) R&D spillovers are expected to be weaker in
countries with higher values for these indices. To
examine whether R&D spillovers are affected by the
power of labour unions in limiting the take-up of
potentially labour-saving technology, further
information on trade union density from the OECD is
included (Union). A further indicator employed is the
OECD indicator of product market regulation (PMR).
The indicator captures the stringency of product
market regulatory policy, with higher values being
associated with policies that are more restrictive to
competition72. A further variable that is included is an
indicator of the strength of Intellectual Property
Rights (IPR). IPRs are a policy tool to encourage
innovative activities. By preventing the copying and
imitation of a patent, IPRs may reduce technology
diffusion. However, since information on patents is
71
See http://www.barrolee.com/. These data have been used as a
measure of absorptive capacity in similar studies (see for
example Falvey et al. 2007).
72
This indicator ranges on a scale from 0 to 6. The data is
available at the country-level only and for three years (i.e.
1998, 2003, 2008). Missing years are imputed using linear
interpolation. For further details see Wlfl et al. (2009).
77
made public, stronger IPRs may encourage
technology diffusion (Breitwieser and Foster 2012).
The index of IPRs used is developed by Ginarte and
Park (1997) and updated by Park (2008). This index
uses information on the coverage of patents,
membership in international treaties, enforcement
mechanisms, and restrictions on patent rights and
duration, with higher numbers indicating stronger
protection73.
Finally, information from the Heritage Foundations
Index of Economic Freedom is used for additional
variables. In particular, the sub-indices on investment
freedom (invest) and financial freedom (finance) are
included, where higher numbers imply more
restrictions74.
3.2.1. Empirical model and estimation
To study the impact of foreign R&D content of
intermediates on labour productivity the following
production function is used:
(3.1)
The growth rate of labour productivity is
dependent on the growth rate of foreign R&D
spillovers , the growth rate of the capital-labour
ratio and the initial lagged value of output per
worker to allow for conditional convergence. Further
industry , country and time fixed effects are
included, while is an error term. The inclusion of
fixed effects controls for unobserved heterogeneity
across the respective dimensions. Results are reported
in Table 3.1. The negative and significant coefficient
on initial output per worker confirms the presence of
conditional convergence.
The coefficient on the capital-labour ratio is positive
and significant in all specifications, indicating that
greater capital intensities are associated with higher
labour productivity growth. With respect to the main
variable of interest, the coefficient estimates indicate
that a 1% increase in the growth of foreign R&D
content of intermediates is associated with a higher
growth rate of labour productivity of between 0.15%
and 0.19%. Thus, these results suggest that the R&D
stock of intermediates is positively associated with
output per worker.
73
The index takes on a value between zero and five.
74
The raw data are on a scale of zero to 100, with 100 implying
no restrictions. To be consistent with the other measures of
absorption barriers this variable is redefined to be equal to
. For further details on the
construction of these indicators see the background study.
The second question to address is whether the
relationship between the foreign R&D stock of
intermediates and labour productivity is affected by
the indicators of absorptive capacity and absorption
barriers described above. The econometric strategy
involves estimating a model of the following form:
(3.2)
where is the indicator of absorptive capacity or
absorption barriers and is the indicator function
taking the value one if the term in brackets is true.
The model differs from a standard linear model in
that the elasticity of labour productivity with respect
to foreign R&D (i.e. ) is allowed to differ depending
upon whether absorptive capacity is above or below
some threshold value ( ). In particular, the elasticity
of labour productivity is given by if absorptive
capacity is below (or equal to) the threshold and is
given by if absorptive capacity is above the
threshold. The actual threshold value is calculated
endogenously following Hansen (1996, 1999 and
2000) with significance of the thresholds determined
by bootstrapping (see Annex 2c). When estimating
this model the threshold variable, , is included
linearly. The set of threshold variables capturing
absorptive capacity and absorption barriers also
includes the initial values of labour productivity
( ). This allows for the examination of whether
an indicator of relative backwardness impacts upon
the relationship between foreign R&D and labour
productivity. While being further behind the
technological frontier usually means that there is
more scope for technological catch-up it could also
imply that a country or sector does not have the
ability to make use and benefit from advanced
technology (see Falvey et al. 2007). As such, the
impact of backwardness measures on the relationship
between foreign R&D and labour productivity growth
is ambiguous from a theoretical point of view.
Results from estimating a single threshold are
presented in Table 3.2. where each column presents
the threshold estimates for the variable indicated in
the header line as motivated in the text above.
Coefficients on initial output per worker and the
growth of the capital-labour ratio are consistent with
the results in Table 3.1. In terms of the threshold
results, a variety of outcomes appears. The
backwardness measure shows that the lower the
labour productivity the larger the spillover effects.75
75
Alternatively one might use interaction terms between R&D
and absorptive capacities and barriers. However, the use of the
threshold model rather than interaction terms has a number of
78
The coefficient in the low regime (0.264) is more
than twice the coefficient in the high regime (0.105)
though both are significant indicating that foreign
R&D spillovers appear to be significantly stronger in
countries and industries that are further away from
the frontier.
The indicators of absorptive capacity (i.e. average
years of secondary schooling, Syr, and lnR&D),
produce consistent results. The coefficients indicate
that foreign R&D spillovers are larger in countries
with a higher number of average years of secondary
schooling and in countries and industries which are
more R&D intensive. While the difference in
coefficients (0.11 versus 0.27) in the case of Syr is
significantly different, the differences in the case of
lnR&D (0.15 versus 0.17) are not significant, i.e. the
linear model is preferred.
When labour market indicators are used as a
threshold variable, results vary across the different
indicators. When using indicators of the strength of
regulation on regular contracts and collective
dismissal, spillover effects are larger in the low
regime (i.e. in countries with less regulations). The
coefficient estimates imply that a 1% increase in the
growth of the foreign R&D stock has a 0.13%
increase in labour productivity growth for countries
with a value of the EPR below the threshold and a -
0.001% decrease for countries above the threshold. A
similar change increases labour productivity growth
by 0.21% for countries with EPC below the
threshold, and by just 0.04% for countries above the
threshold. The strength of regulation on temporary
contracts produces the opposite result. In particular,
while a 1% increase in the growth of the foreign
R&D stock is associated with an increase in labour
productivity of 0.24% for countries with EPT above
the threshold, the change for countries below the
threshold is just 0.03%. Finally, when using union
advantages. Firstly, using threshold models doesnt impose a
monotonic change in the effect of the explanatory variable as
the threshold or interaction term increases (i.e. the impact of
the explanatory variable on the dependent variable can switch
signs and change size at different points on the distribution of
the threshold variable). Secondly, the coefficients are easier to
interpret. The impact of the explanatory variable on the
dependent variable is given by a fixed parameter for all
observations within a particular regime. With interaction terms
it is more difficult to identify the overall impact of a change in
the explanatory variable, with researchers often resorting to
graphing the relationship for different values of the
threshold/interaction variable. Thirdly, when the
threshold/interaction variables are bound as in our case (e.g.
between zero and six) the threshold model is less open to
misinterpretation (e.g. extrapolating beyond the range of the
threshold/interaction variable).
density as threshold variable results show that foreign
R&D spillovers are larger in the low union density
regime. A 1% increase in the growth of foreign R&D
is associated with a 0.19% increase in labour
productivity growth in the low regime, and a 0.03%
increase in the high regime.
foreign R&D content of intermediates is positively
associated with labour productivity growth. However,
the size of these spillovers depends on absorptive
capacities and barriers; countries and industries
further behind the technological frontier enjoy
stronger foreign R&D spillovers, in line with Falvey
et al. (2007). The results also support Falvey et al

Table 3.1. Foreign R&D and labour productivity growth

(1) (2) (3)

-0.0104*** -0.0106*** -0.0142***

(0.000741) (0.000797) (0.00289)
0.482*** 0.422*** 0.465***
(0.0278) (0.0301) (0.0334)
0.190*** 0.176*** 0.150***
(0.0180) (0.0192) (0.0202)
Time F.E. No Yes Yes
Country F.E. No No Yes
Industry F.E. No No Yes

Observations 15,850 15,850 15,850

R-squared 0.372 0.419 0.455
F-stat 289.2*** 338.2*** 87.04***
Notes: ***, **, * significant at 1, 5 and 10%. Robust standard errors in parentheses.

In terms of the remaining indicators, one finds that, in
the cases of , and , the
relationship between foreign R&D growth and labour
productivity growth is stronger in the high regime,
that is, in the regime with more stringent product
market, investment and financial regulation. In the
case of the coefficient in the low regime is
negative and significant. For the difference in
the coefficients on the foreign R&D variable between
the two regimes is relatively small though still
significantly so (0.149 versus 0.172), while for
the differences are much larger (0.08 versus
0.237). Though this might be an unexpected result, it
should be noted that these indicators could also
reflect institutional quality in a broader sense. As
countries with higher institutional quality might
attract more R&D intensive firms or have tighter
cooperation in R&D activities, etc. these results
would be in line with the literature stating that the
quality of institutions matters.
Summarising, the results confirm those from the
simple model reported in Table 3.1. whereby the
(2007) as well as Crespo-Cuaresma et al (2008) in
finding that foreign R&D spillovers are stronger in
countries with greater absorptive capacity (as
measured by average years of secondary schooling
and R&D spending). In terms of absorption barriers,
the results are mixed. With the exception of
regulations on temporary workers, stronger labour
market regulation and greater union density are
associated with lower foreign R&D spillovers,
consistently with Crespo-Cuaresma et al. (2008) and
Parente and Prescott (1994, 1999 and 2003).
Concerning the other absorption barriers related to
product market, financial and investment regulation
there is no evidence that lower anti-competitive
barriers encourage foreign R&D spillovers. Indeed,
the reverse appears to hold though these indicators
might reflect the overall institutional quality, which is
conducive to growth. Finally, one finds that stronger
levels of IPR protection can limit the extent of foreign
R&D spillovers, possibly by limiting the ability to
replicate and borrow technology from abroad.

79
Table 3.2. Single threshold results
Threshold (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
variable Syr EPR EPT EPC Union PMR IPR Invest Finance

-0.011*** -0.014*** -0.014*** -0.00488 -0.00430 -0.00452 -0.00488 -0.00428 -0.013*** -0.0143*** -0.0142***
(0.00226) (0.00225) (0.00226) (0.00305) (0.00304) (0.00325) (0.00300) (0.00292) (0.00234) (0.00226) (0.00224)
0.457*** 0.465*** 0.465*** 0.629*** 0.632*** 0.664*** 0.631*** 0.621*** 0.467*** 0.465*** 0.454***
(0.00587) (0.00586) (0.00587) (0.0104) (0.0104) (0.0115) (0.00993) (0.00999) (0.00611) (0.00589) (0.00588)
0.264*** 0.111*** 0.149*** 0.126*** 0.0307*** 0.208*** 0.191*** -0.0291** 0.211*** 0.139*** 0.0807***
(0.0107) (0.00766) (0.00631) (0.0110) (0.00999) (0.0239) (0.0129) (0.0121) (0.00759) (0.00732) (0.00792)
0.105*** 0.212*** 0.174*** -0.000967 0.241*** 0.0357*** 0.0339*** 0.156*** -0.057*** 0.172*** 0.237***
(0.00705) (0.00952) (0.0226) (0.0145) (0.0182) (0.0104) (0.00963) (0.0108) (0.0127) (0.0103) (0.00904)
0.00358 9.47e-05 -0.00574 -0.026*** -0.072*** 0.000722 -0.00141 -0.00102 -0.00041*** 0.000384***
(0.00754) (0.000390) (0.0131) (0.00347) (0.0111) (0.000475) (0.00830) (0.00376) (0.000133) (0.000126)

Threshold 1.566 3.958 12.270 2.470 3.444 1.959 16.498 1.737 4.180 49.506 32.222
Percentile 21 66 50 64 79 13 18 47 70 90 46
P-value 0.000*** 0.000*** 0.297 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.008*** 0.000***
Observations 15,850 15,850 15,850 9,559 9,559 8,061 10,372 9,742 14,200 15,850 15,850
R-squared 0.461 0.458 0.455 0.448 0.453 0.471 0.461 0.479 0.468 0.455 0.462
F-stat 158.4*** 154.6*** 153.0*** 98.60*** 100.5*** 93.59*** 120.8*** 119.9*** 151.5*** 153.3*** 157.7***
Notes: ***, **, * significant at 1, 5 and 10%. Robust standard errors in parentheses. All models include unreported time, industry and country fixed effects

80
3.3. EFFICIENCY ANALYSIS AT THE INDUSTRY
LEVEL
This section provides an analysis of the determinants
of technical efficiency in Europe, the US and Japan,
using stochastic frontier analysis (SFA). Technical
efficiency in this study refers to the ability of a
firm/industry to achieve the maximum output using
the set of available resources. The growth accounting
and the regression analysis framework used in the
previous sections are based on the assumption that all
resources, i.e. capital and labour inputs, are fully
utilised and therefore it cannot account for changes in
productivity originating from efficiency
improvements. Technical efficiency analysis relaxes
this assumption and assumes that only the top
performing industry is able to use resources in the
most efficient way. The other industries will lie below
the frontier and the distance to the frontier output
defines the efficiency gap. Identifying which
industry/country is at the frontier and how efficiency
levels have changed over time is important to direct
policies towards the correct tool to promote
performance. In fact, the best performing industry in
terms of productivity might not be the most efficient
one, and higher productivity could be achieved by
improving the allocation and usage of the available
resources. On the other hand, a highly efficient
industry might not be the most productive because of,
for example, low investments in strategic assets such
as ICT and R&D capital; in this case, policies should
be directed towards promotion of investments.
The most intuitive way of understanding frontier
analysis is to assume that the actual output produced
can be lower than the maximum output, given the
level of available resources. By defining actual output
in industry i at time t as YAit and the maximum output
as YFit, technical efficiency can be expressed as:
(3.3)
Efficiency levels in each industry range between 0
and 1, with higher scores indicating higher efficiency.
The derivation of technical efficiency levels requires
the estimation of a production function where output,
measured by value added, is produced with a
combination of inputs. The most basic model includes
the total number of hours worked (H) and total capital
(K) as factor inputs. Below, a more extended
specification will also be considered that accounts for
different types of capital (ICT and non-ICT capital)
and intangible assets (labour quality and R&D
capital).
The analysis is carried out using industry-level data,
extracted from the EUKLEMS database. The total
sample includes 16 countries. Of these, 14 are
European (AT, BE, CZ, DE, DK, ES, FI, FR, HU, IE,
IT NL, SE and UK); the other two being Japan and
the US. For each country, data are available for 21
81
industries, including manufacturing and services76.
The analysis is conducted between 1995 and 2007
outlining industry performance in the pre-financial
crisis period. The exclusion of the downturn is
motivated not only by data availability, but also by
the consideration that technical efficiency relates to
the structure of the industry, and this is less likely to
be affected by cyclical factors or exogenous shocks.
Results from the estimation of a frontier production
function, expressed in log-levels, are presented in
Table 3.377. The first column (1) displays estimates
for total capital services, while column (2)
distinguishes between ICT and non-ICT capital
services. Results are robust across the two
specifications and, with the exception of ICT capital,
all coefficient estimates are positive and statistically
significant. They are consistent with prior knowledge
of factor shares. Human capital, measured by the
labour quality variable, has a positive, albeit small,
effect on productivity. In column (2), the impact of
ICT is positive with an elasticity of 0.04%, which is
consistent with the existing evidence (Kretschsmer
2012). However, the coefficient is not statistically
significant at conventional levels. It is possible that
the model specification needs to account for
additional complementary assets (Diedrick et al.
2003). In fact, when R&D is included, the
significance of the ICT variable improves as in
column (4). Another possible explanation is that the
impact of ICT is highly heterogeneous across
countries and industries, and its effect is likely to be
higher in the most ICT intensive users. More
importantly, further developments of this analysis
will show that ICT exerts an indirect effect on
productivity, via the reduction of technical
inefficiencies.
Results in Table 3.3. columns (3) and (4), confirm the
importance of R&D in increasing productivity,
consistently with the reference literature, where this
value generally ranges between 0.04 and 0.18
(Griliches and Mairesse 1984; Kumbhakar et al.
2010; Bloom et al. 2013).
76
Industries included in the analysis are: Food and Beverages,
Textile and Leather, Wood & Cork, Pulp, Paper and Printing,
Coke, refined petroleum and nuclear fuel, Chemicals, Rubber
and Plastic, Other non-metallic minerals, Basic metals,
fabricated metal products, Machinery NEC, Electrical
Equipment, Transport Equipment, Manufacturing NEC,
Transport and Storage, Post and Telecommunication, Business
Services, Electricity, Gas and Water, Construction, Wholesale
and Retail, Financial Intermediation, Other Community and
Social Services. This classification is based on NACE Rev. 1
and differs from the one used in the Section 3.1.3.
77
The estimation of the production function in the panel
dimension requires the introduction of fixed effects to control
for cross-sectional time-invariant heterogeneity. Along with
country- and industry-specific intercepts, the specification also
includes a set of time dummies to control for unknown or
unobserved factors that are likely to affect all industries at
different points in time.
Table 3.3. Estimation of stochastic frontier production function
Dependent variable: Value added
(1) (2) (3) (4)

Total number of hours worked 0.790*** 0.755*** 0.632*** 0.612***

(0.031) (0.031) (0.036) (0.037)

Tangible assets
Total capital 0.351*** 0.446***
(0.024) (0.028)

Non-ICT capital 0.394*** 0.445***

(0.024) (0.029)
ICT capital 0.019 0.026
(0.017) (0.020)
Intangible assets
Labour quality 0.011*** 0.010*** 0.008*** 0.007***
(0.002) (0.002) (0.002) (0.002)
R&D capital 0.061*** 0.061***
(0.008) (0.008)
Constant 1.418*** 1.154*** 1.246*** 1.248***
(0.168) (0.171) (0.189) (0.189)

Gamma 0.401 0.343 0.394 0.433

Likelihood ratio test 11.025 5.176 7.434 9.663
(P-value) (0.000) (0.011) (0.000) (0.000)
Observations 4532 4519 3650 3488
Notes: ***, **, * significant at 1, 5 and 10%. Gamma is the proportion of the total error variance due to inefficiencies. The Likelihood
ratio test is a test of the null hypothesis that there are no technical inefficiencies in production.
Source: EUKLEMS and OECD ANBERD

The SFA modelling framework allows the derivation

The inclusion of R&D does not significantly affect
the coefficient estimates for labour quality, and
generates only a marginal increase in the effect of
total capital see column (3) and non-ICT capital in
column (4)78. This is a consequence of the
complementary relationship between R&D and
capital assets79.
78
Diagnostic statistics are presented at the bottom of Table 3.3.
The Gamma parameter measures how important inefficiencies
are in each model. A value of 1 indicates that all deviations
from the frontier are due to inefficiency, while a value of 0
implies that there are no inefficiencies; in the latter case, SFA
does not provide any additional information compared to OLS.
In this study, the Gamma parameter is approximately equal to
0.4 meaning that inefficiencies are important and explain 40%
of the total residual variation. The presence of inefficiencies is
also assessed via the Likelihood Ratio test, which confirms
that this component is statistically significant.
79
The impact of the total number of hours worked is
significantly lower in regressions 3 and 4, compared to the first
two columns of Table 3.3. This is related to the fact that a large
of technical efficiency (TE) for each industry/time
period. Estimates of technical efficiency can be
derived from any of the specifications presented in
Table 3.3. , hence a choice needs to be made to carry
out the analysis. The last row of the table shows that
the number of observations drops substantially when
including R&D, as information on this asset is
missing for several service industries in various
countries. Given that the main objective of this
section is to analyse efficiency trends across the full
spectrum of manufacturing and services, the
specification in column (2) is used to derive technical
efficiency scores80.
proportion of R&D costs is composed of the wages of
employees involved in R&D activities. This 'double counting'
is a well-known phenomenon in productivity studies
(Schankerman1981, Guellec and van Pottelsberghe 2004).
80
The correlation of TE scores arising from the four
specifications is very high, ranging between 0.97 and 0.99.

82
Figure 3.7. Average technical efficiency (TE) in the EU,
has been increasing over time in the most innovative
US, and Japan sectors, namely ICT producing and high-tech
industries. This suggests that increases in productivity
0.800 went hand-in-hand with increases in efficiency until
0.790 2007. The efficiency in the ICT-producing sector
0.780
increased by 5%, from 0.75 in 1995 to 0.80 in 2007,
TECHNICAL EFFICIENCY

while improvements were more moderate in high-

0.770
tech industries (approximately 2%).
0.760
Panel B of Figure 3.9. focuses on services,
0.750
distinguishing between knowledge-intensive and non-
0.740
knowledge intensive industries. While the overall
0.730 performance of the tertiary sector has been declining
0.720 over time, this figure reveals that the average picture
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 is influenced by the dynamics of low knowledge-
EU US JP
intensive services, as they are characterised by a
steady decrease in TE. On the other hand, the most
Source: EUKLEMS Database and authors computations.
knowledge intensive industries, after a dip in
Figure 3.7. presents average efficiency scores for the Figure 3.8. Technical efficiency
EU, the US and Japan. Although actual efficiency Panel A. Manufacturing and services
levels do not differ greatly across economies, their 0.790
variation over time shows some interesting patterns. 0.785
In the mid-1990s, the US had lower efficiency levels 0.780
than Japan and Europe, but then TE increased rapidly
TECHNICAL EFFICIENCY

0.775

placing the US at the frontier since 2002. Existing 0.770

evidence dates the resurgence of US productivity to 0.765

0.760
1995, around seven years before the aggregate
0.755
increase in efficiency. This difference can be 0.750
explained by the presence of lags in the full 0.745
implementation of the new technology and the 0.740
reorganisation of production, emphasised in the 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Manufacturing Services
General Purpose Technology (GPT) literature
(Hornstein and Krusell 1996, Aghion 2002). While
the existing evidence mainly refers to the direct Panel B. ICT producing and using industries
impact of ICT on productivity, here the analysis
provides new results supporting the GPT nature of
new digital technologies. From 2002 to 2005 the
efficiency gap between the US and the other two
economies widened. However, from 2005 efficiency
levels fell in the US, while the EU trend remains
virtually unchanged. Japan was the frontier country in
1995 but its efficiency declined from 1996, and since
2000 its efficiency has been below the European
average. These patterns are not dissimilar from trends
in TFP levels discussed in Jorgenson and Nomura
(2007). Similarly to the US, in Japan changes in
efficiency follow changes in productivity with a lag Source: EUKLEMS Database and authors computations.
of about five years.
A look at average TE in selected groups of industries efficiency in 2000, performed relatively well, with
provides insights on the ones performing better. Panel increasing efficiency throughout the period.
A of Figure 3.8. presents mean efficiency trends in
manufacturing and services. This shows that services
have experienced a declining efficiency performance
over time, while in manufacturing efficiency has
remained fairly stable over the period. Figure 3.8.
panel B and Figure 3.9. panel A show that efficiency

Hence, the exclusion of R&D does not affect the estimation of

TE.

83

Figure 3.9. Technical efficiency in high-tech and low-
tech industries
Panel A. High-tech and low-tech
Panel B. Knowledge and non-knowledge intensive
services
Source: EUKLEMS Database and authors computations.
3.3.1. Reducing efficiency gaps: discussion of the
main determinants
Understanding why industries vary in the extent to
which they use resources effectively, and what
policies might be more suitable to foster efficiency
performance, requires the extension of frontier
analysis to account for the factors that might cause
industries to fall below the frontier and therefore
widen efficiency gaps. This study focuses on the role
played by ICT capital and the business environment
where industries operate. The assessment of the
impact of these factors on technical efficiency is
obtained by the empirical estimation of the following
relationship:
(3.4)
is a simple time trend that captures how an
efficiency gap is determined by exogenous
84
technological changes81. it is the error term. The
inclusion of a large number of indicators naturally
causes co-linearity problems, hence the researcher
needs to deal with the trade-off between efficiency of
the estimator (which is reduced in the presence of
collinearity) and omitted variable bias. To address
this issue, the estimation will sequentially include
different indicators, checking for the presence of ICT
and the degree of competitiveness in all
specifications82. The decision to include these two
factors is driven by the existing evidence that
suggests the presence of complementarities between,
for example, product market regulation and
employment protection legislations (Griffith et al.
2007, Fiori et al. 2012)83. The analysis uses a wide
range of indicators that have been rescaled so that all
vary between 0 and 1, with larger values indicating
more stringent regulation.
Table 3.4. to Table 3.6 present the sign of the impact
for each of the factors affecting technical efficiency,
and the related statistical significance. Table 3.4.
presents a summary of the results based on a
specification that includes ICT and a set of indicators
capturing the degree of market competitiveness. The
latter includes the Upstream Regulation Index (RI),
which assesses the impact of anti-competitive
legislation in the tertiary sector on the performance of
downstream sectors that use services as a production
input (Conway et al. 2006); Enforcing Contract Time
(ECT), which is based on the number of days to
enforce a contract in each country (World Bank
2012); and two alternative measures of
competitiveness, the Herfindal index and the degree
of industry fragmentation84. The coefficient for ICT is
negative and statistically significant in all
specifications, indicating that this asset plays a very
important role in lowering inefficiencies in the use of
resources. Although ICT did not have a significant
effect in the estimation of the production function, it
plays an important role in reducing efficiency gaps,
which will also affect productivity but in an indirect
way. This result is particularly interesting when
compared with the existing industry-level evidence,
which usually fails to find significant effects of ICT
on TFP growth (Stiroh 2002, Basu et al. 2004,
Acharya and Basu 2010). This implies that
81
The estimation of the efficiency gap is carried out
simultaneously with the estimation of the productivity frontier,
using Maximum Likelihood methods. This one-step procedure
82
guarantees consistency in the coefficient estimates.
The authors also tried to include intangible assets (R&D and
labour quality) as determinants of both productivity and
efficiency. However, these estimates were highly unstable,
hence such factors were only included in the specification of
83
the production function (see Table 3.3).
The impact of the latter factor is always accounted for with the
use of the upstream Product Market Regulation index (RI),
unless otherwise specified.
84
The Herfindal index and the indicator of industry
fragmentation are derived using information from the
Amadeus database, made available via EUKLEMS.
distinguishing between TFP and TE can provide
important insights on the role of ICT. Up to now, this
issue has been unexplored by the economic literature.
Consistently with the existing evidence on the link
between the lack of competition and productivity
(Buccirossi et al. 2012, Conway et al. 2006), Table
3.4. indicates that higher values of upstream
regulation significantly increase the efficiency gap. In
other words, administrative restrictions on
competition in the services market have widespread
negative effects on production efficiency, well
beyond the tertiary sector. This effect is robust to the
use of alternative variables that describe the degree of
competitiveness of the market, such as the Enforcing
Contract Time (ECT), the Herfindal index and
industry fragmentation. Note that the sign of the latter
is negative as higher values indicate higher
fragmentation, which is associated with more
competition. Hence, results in Table 3.4. provide
strong support for the hypothesis that a more
competitive business environment reduces the
efficiency gap.
Table 3.4. Reducing the efficiency gap: the role of product market competition
(1)
ICT capital -0.623***
Upstream regulation (RI) 3.081***
Enforcing contract time
Herfindal
Industry fragmentation
Observations 3648
Note: ***, **, * significant at 1, 5 and 10%. All specifications include a time trend. A negative sign implies that the variable decreases the
efficiency gap
Table 3.5. reports results based on a specification that
includes indicators of employment protection
legislation. Three of these (EPR, EPT, EPC) are
described in Section 3.2. The EPL burden indicator is
an industry level variable based on the notion that
EPL, although it applies uniformly to all industries
within a country, is more binding in those industries
that rely on lay-offs than in industries characterised
by a higher degree of voluntary turnover (Bassanini et
al. 2009). NMW stands for National Minimum Wage
and it is only available for those countries where
NMW is prescribed by law (Visser 2011).
Considering the role of NMW is interesting in the
light of the political and economic debate concerning
its effect on employment outcomes (Card and
Krueger 1995, Neumark and Washer 2008). Results
in Table 3.4. show that more stringent EPL for
regular workers and collective dismissals
significantly increases the efficiency gap. On the
other hand, regulation on temporary workers has the
opposite effect, indicating that stronger protection on
this kind of contracts decreases inefficiencies. This
finding is not unexpected as existing evidence shows
that the legal discipline on regular and temporary
workers can have opposite effects on performance.
For example, Damiani et al. (2013) document that
deregulation of temporary contracts negatively
influences TFP growth in European industries. These
findings suggest that excessive flexibility in the use
of temporary workers leads firms to use this category
of workers to buffer cyclical demand movements
(Gordon 2011), rather than attempting to find the
most efficient way to combine factor inputs85. The
introduction of minimum wage legislation also
constrains the efficient use of labour input by
increasing the efficiency gap, through reducing
competitiveness in the labour market, a claim that is
frequently used by those scholars who opose the
introduction of the minimum wage (Currie and
Fallick 1996)86.
The final set of indicators accounts for the effect of
financial market regulation, property right protection
and regulation of FDI on the efficiency gap. Financial
market regulation is measured using three indicators:
the financial reform index, constructed by combining
(2) (3) (4)
-0.636*** -0.693*** -0.004
3.111***
0.639***
0.747**
-1.097***
3648 2454 2256
liberalisation scores on seven different areas of the
financial market (Abiad et al. 2008); the financial
freedom index, as defined in Section 3.2; and the ratio
of product market capitalisation over GDP (Beck et
al. 2009). The indicator of property rights regulation
is defined in Section 3.2. The FDI regulation index
summarises information on four forms of legal
intervention (equity restrictions, screening and
approval requirements, restrictions on foreign key
personnel and other operational restrictions; see
Kalinova et al. 2010). Results including these
indicators are presented in Table 3.6. Both the
financial reform and the financial freedom indicators
consistently show that lower levels of regulation in
this market increase the
85
This result is also consistent with those in section 3.3 of this
chapter, where countries with higher employment protection
for temporary workers enjoyed higher spillovers from foreign
R&D stocks.
86
In the basic textbook model of labour demand an increase in
the minimum wage reduces the employment in the covered
sectors of those workers whose wage rates would otherwise
fall below the minimum (Currie and Fallick 1996; p. 405).
85
Table 3.5. Reducing the efficiency gap: the role of employment protection legislation
(1) (2) (3)
ICT -0.574*** -0.559*** -0.557***
Upstream regulation (RI) 4.304*** 2.926*** 3.141***
EPL burden indicator 0.142***
EPL regular 4.351*** 3.513***
EPL temporary -2.173*** -3.007***
EPL collective dismissal 2.716**
NMW 1.372***

Observations 3146 3648 3021

Note: ***, **, * significant at 1, 5 and 10%. All specifications include a time trend.
openness is not a major issue (note that the coefficient
efficiency gap. Conversely, access to alternative
is only significant at the 10% significance level).
sources of finance, like the bond market, significantly
Additionally, next to the literature which emphasises
improves TE performance. This follows the main
the importance of trade openness for growth, there are
results of the literature, where a positive relationship
also contributions that support the positive role of
between financial development and growth is usually
protectionist measures. For example, estimates in
found (Rajan and Zingales 1998, Maskus et al. 2012),
Yanikkaya (2003) predict a positive and significant
but where it is also suggested that excessive financial
relationship between trade barriers and growth.
liberalisation may be detrimental for performance
Although these results are driven by developing
when it discourages savings or triggers financial
countries, they nevertheless imply that the
instability (Ang 2011, Ang and Madsen 2012). The
relationship between trade and growth is quite
Table 3.6. Reducing the efficiency gap: financial regulation, intellectual property rights protection, openness
(1) (2) (3) (4)
ICT capital -0.626*** -0.545*** -0.496*** -0.502***

Upstream Regulation (RI) 3.057*** 2.441*** 2.204*** 2.255***

Financial Reform Index -1.451 -1.849**

Private Bond Mkt Cap/GDP -0.307*** -0.317*** -0.288***

Financial Freedom -2.650*** -2.776***

IPR -0.080

Regulation of FDI -1.303*

Observations 3648 3648 3648 3648

Note: ***, **, * significant at 1, 5 and 10%. All specifications include a time trend.

recent financial crisis has shown that excessive complex. Moreover, the present analysis deals with
freedom in the financial market can have catastrophic the specific issue of technical efficiency and it is
consequences. possible that the impact of FDI openness on growth
differs from the effect of this factor on technical
Results also show that increasing protection of
efficiency. Additionally, product market and labour
intellectual property reduces the efficiency gap,
market regulations might prevent or delay the
although the effect is not statistically significant. This
necessary adjustments in the combination of inputs,
suggests that property rights regulations might not be
which would allow countries to fully benefit from
relevant for efficiency improvements. The measure of
globalisation. Hence, it is possible that increasing
openness to external markets, summarised by the FDI
international openness may lead to higher levels of
regulation index, shows that stricter FDI rules
production efficiency only over a relatively long time
decrease the efficiency gap. Although this goes
horizon. Further investigation of this issue goes
against the extensive literature on the positive relation
beyond the scope of the present analysis but suggests
between trade openness and growth, there are several
an interesting development for future research.
reasons that might explain this result in the sample
used in this study. Firstly, the countries considered
have very low levels of regulations, hence trade

86
3.4. EU PRODUCTIVITY PERFORMANCE AT THE
FIRM LEVEL: EVIDENCE FROM EU-EFIGE
SURVEY ON MANUFACTURING FIRMS
This section complements the industry level analysis
with evidence based on a newly available firm level
dataset (EU-EFIGE) which collects information on
14,759 manufacturing firms across seven EU
countries (Austria, France, Germany, Hungary, Italy,
Spain and the UK)87, over the period 2007-2009. The
analysis focuses on how the financial crisis has
affected performance and innovation strategies at the
micro level. The focus is on productivity (TFP) rather
than efficiency, as data constraints prevent the use of
Stochastic Frontier methods.
Figure 3.10. TFP growth 2001-07 and 2008-09 (by
country, size class and Pavitt)
The financial crisis of 2008-2009 was a watershed for
the European Union as it widened growth disparities
among the Member States.
dynamics. Large companies (over 249 employees)
outperformed other types of firms in all productions88.
After a period of moderately positive rates of TFP
growth, medium-sized firms (50-249 employees)
faced a severe drop during the crisis. A relevant
exception is in the science-based firms whose
productivity growth was positive although slightly
reduced during the crisis with respect to the early
2000s. The collapse of the market in 2008-2009
severely hit traditional industries, which experienced
a fall in TFP levels of almost 30%, especially because
of small firms performance.

Figure 3.10. (Panel A) describes TFP changes

between 2008 and 2009 compared to the average
annual rate of growth in the pre-crisis period (2001-
2007). The average rate of TFP growth was negative
in the early 2000s, i.e. before the crisis, in France,
Italy and Spain. With the downturn, productivity
performance worsened in all countries, but remained
positive in Austria, Germany and Hungary. The UK
is the country where the deceleration in TFP was
most dramatic between the two sub-periods, falling
from a positive rate of 16.5% to -15%. In absolute
values, Italian firms registered the worst rates of TFP
change in the downturn (-29%). There are some
important regularities in productivity dynamics in
firm-level performance. On average, those firms
performing better in terms of TFP growth before the
downturn, as measured by the average rate of change
between 2001 and 2007, also presented higher rates
of productivity growth during the period 2008-2009.

Panel B in Figure 3.10. provides a breakdown of TFP

performance, distinguishing among size classes and
industry categories. Confidentiality issues prevent the
use of detailed industry information, hence the
analysis will follow the Pavitt (1984) taxonomy to
control for major industry characteristics.

Small-sized companies (fewer than 50 employees)

were diffusely characterised by negative productivity

Note: S=small firms. M=medium firms. L=large firm, Source:

EFIGE dataset
87
The number of firms varies across countries, with
approximately 3,000 firms in France, Germany, Italy and
Spain, 2067 firms in the UK, 443 in Austria and 488 in
88
Hungary. The sample was originally designed to be Small firms (less than 50 employees) make up 73% of the
representative of the manufacturing sector and, to this aim, was overall sample, medium firms (between 50 and 249
stratified along three dimensions: industries (11 NACE-CLIO employees) account for 20% and large firms (more than 249
industry codes), regions (at the NUTS-1 level of aggregation) employees) accounts for 7% of the sample. This implies that
and size class (10-19; 20-49; 50-250; more than 250 large firms are over-represented, due to their relevance in
employees). The dataset does not provide information on firms aggregate competitiveness dynamics (Altomonte and
exiting the market due to the crisis. Aquilante 2012; p. 5).

87
Table 3.7. Percentage of firms experiencing a turnover reduction in 2009 compared to 2008 (by size class)
AUT FRA GER HUN ITA SPA UK Total
Small firms 60.3 69.6 61.6 75.4 74.3 82.2 66.7 71.5
Medium firms 72.2 73.4 68.9 82.2 80.0 80.8 64.7 72.9
Large firms 67.4 73.8 56.9 68.9 80.0 84.3 65.7 69.7
Total 63.7 70.7 63.1 76.4 75.4 82.1 66.1 71.7

Source: EFIGE dataset

Table 3.8. Percentage of firms experiencing a turnover reduction in 2009 compared to 2008 (by Pavitt groups)
AUT FRA GER HUN ITA SPA UK Total
Economies of scale 66.7 71.9 64.5 77.4 75.3 87.4 68.4 73.0
High-tech 65.2 49.2 48.0 73.3 54.5 69.8 49.5 53.9
Specialised 74.2 72.2 71.9 76.4 77.5 83.9 68.1 74.8
Traditional 61.4 71.9 59.7 75.7 76.5 80.8 65.1 71.9

Source: EFIGE dataset

Figure 3.11. Firms reducing product/process

innovation in 2009 (by country, size class and Pavitt)
of EU firms in terms of turnover fall. In Europe, sales
declined in 72% of the companies in 2009 compared
to the pre-crisis values. This share is considerably
higher for Spain (82%), Hungary and Italy (around
75%). In Germany, the percentage of the sample
facing a turnover fall is 63%. As Table 3.7. shows,
the financial turmoil caused a real downturn that was
very pervasive involving all types of firms.

The industry breakdown provided by Table 3.8.

illustrates that specialized suppliers were hit by the
crisis as 75% of companies experienced a decrease in
sales. From this perspective, the crisis looks more
severe in Spain, Italy, and Hungary where the
proportion of firms facing a turnover reduction was
above the EU average. Scale-intensive firms were
also considerably affected by the crisis, particularly in
Spain where almost 90% of the total sample reduced
sales. On the other hand, the fall of turnover was less
pervasive among high-tech firms.

Rates of investment fluctuate remarkably along the

Source: EFIGE datase
Firm-level productivity is pro-cyclical as, in the short
term, it reflects shocks to demand conditions.
Therefore, to better understand cross-country
differentials in TFP dynamics, it is necessary to look
at how the 2008-2009 crisis affected the performance
business cycle, reflecting firms' expectations of future
sales and profitability. In the recent downturn, the fall
in investment was exacerbated by the credit crunch.
Along with the intensity of the business cycle, cross-
country differentials in firms' capital formation reflect
disparities in the structure of the domestic financial
systems. In Europe, 43% of firms reduced planned
investment in equipment between 2008 and 2009
(ICT and non-ICT assets). This proportion rises with
firm size, from 42% for small firms to 47.6% of large
firms (Table 3.9. ). The greater sensitivity of
investment in large companies probably depends on a
wider exposure to the international market, and
therefore to the collapse of foreign demand.

88
Table 3.9. Percentage of firms reducing investment in 2009 compared to 2008 (by size class)
AUT FRA GER HUN ITA SPA UK Total
Small firms 32.3 55.1 26.8 46.4 35.8 52.0 39.5 42.0
Medium firms 45.5 54.9 36.0 41.5 36.0 55.5 43.2 44.2
Large firms 40.6 62.6 31.5 60.5 50.4 58.6 43.0 47.6
Total 35.8 55.7 29.8 46.6 36.7 52.9 40.6 42.9
Source: EFIGE dataset

Table 3.10. Percentage of firms experiencing a turnover reduction in 2009 compared to 2008 (by Pavitt groups)
AUT FRA GER HUN ITA SPA UK Total
Economies of scale 37.8 52.8 29.1 46.6 38.2 56.7 40.8 43.0
High-tech 21.1 43.7 22.7 41.7 29.4 42.4 24.7 31.3
Specialised 41.7 55.0 37.4 39.3 40.2 57.8 40.9 44.9
Traditional 37.2 58.8 26.7 50.6 35.6 51.1 41.9 43.2
Source: EFIGE dataset

However, there are big differences among countries,
as large firms performed relatively better in Germany,
Austria and, to a lesser extent, in the UK.
The decrease in investment was quite diffuse
throughout the economy in France and Spain, where
between 55% and 60% of the sample reduced their
commitments. Conversely, the effect of the credit
crunch and the demand fall appears less severe in
Austria, Germany and Italy, where only 30-35% of
manufacturing firms cut investment plans. The
analysis of investment dynamics following the Pavitt
groupings provides additional insights (Table 3.10).
Although the overall average is similar for traditional,
specialised suppliers and scale intensive industries,
there is large heterogeneity across countries.
Supplier-dominated (traditional) firms performed
relatively better in Germany, Italy and Austria, but
struggled in France. Among high-tech firms, Austria
is the country with the smallest decrease in
investments, followed by Germany, the UK and Italy.
Spain and France show a parallel performance in
specialised providers and scale- intensive firms.
Innovation performance of EU firms is examined
using a large spectrum of indicators. The most
comprehensive measure at hand from the EU-EFIGE
dataset is the percentage of companies introducing a
process or/and product innovation between 2007 and
2009 (Table 3.11.).
In Europe, 65% of companies were engaged in such
activities. The breakdown is 61% for small firms,
73% for the medium-sized firms and 79% for the
largest ones. Austria leads in terms of the proportion
of firms involved in innovation, followed by Spain
and Italy. It is well known that this kind of qualitative
indicator is more suited to describe the innovative
capacity in less technologically advanced production,
which explains why Germany is at the bottom of the
ranking. Germany recovers in the ranking when
looking at proxies for formal innovation such as the
proportion of R&D-performing firms, R&D intensity
over sales and patenting.
On average, one out of two EU firms declares that it
carries out R&D projects. Again, Austria shows the
largest proportion of innovators, closely followed by
Germany and Italy. Engagement in formal research is
rather low in Spain and in Hungary89.
The financial crisis and the consequent downturn
caused a drop in firm demand and severely clouded
expectations of future sales. These issues, combined
with tighter credit conditions, led one third of EU
firms to postpone their programmes for product
or/and process innovation (Figure 3.11 Panel A). The
share of firms reducing their engagement in
innovation activities reaches 50% among Spanish
SMEs. A larger heterogeneity emerges when looking
at how firms changed their innovation programmes
because of the deepening crisis across industry
groupings (Figure 3.11, Panel B). Half of the Spanish
firms postponed product/process innovation; this
proportion is similar among Pavitt categories. Apart
from Spain, innovation activities of high-tech
companies were less affected by the turmoil.
89
For a comparison of the EU with non-EU R&D performance
see Moncado-Patern-Castello et al. (2010). This study shows
that the US has a stronger sectoral specialisation in the R&D
intensive (especially ICT related) sectors, compared to the EU.
Furthermore, the population of R&D investing firms within
these sectors is relatively larger.

89
Table 3.11. Firm innovation performance in 2007-2009: summary of results (% of total)
AUT FRA GER HUN ITA SPA UK Total

Product/process innovation 75.9 56.3 64.6 55.7 67.5 69.6 67.3 64.9
Reduction product/process innovation (in 2009) 29.4 30.2 30.4 34.8 35.8 50.1 30.3 35.4
Doing R&D 55.5 50.7 54.6 26.8 55.0 46.0 53.2 51.2
R&D intensity on turnover 6.5 6.2 7.8 5.7 7.3 7.1 6.9 7.0
Patent an innovation 19.4 11.7 15.8 4.3 14.2 11.2 14.0 13.2
Source: EFIGE dataset

intangible factors concur to build the knowledge

3.4.1. Analysis of the productivity effect of the
crisis: econometric evidence
This section investigates how the financial crisis
affected firms productivity performance by
developing an empirical model where the rate of
growth of TFP between 2008 and 2009
( is explained by a large set of
company characteristics, all defined as dummy
variables90. The analysis is based on the following
specification:
(3.5)
To allow for the dynamic profile of productivity
performance, the level of TFP in 2008 is included
among the regressors. A negative coefficient on this
variable would indicate the presence of a catch-up
effect, whereby lower productivity firms fill the gap
with the best performing ones. The first set of dummy
variables (CRISIS in eq. 3.5) seeks to check for the
effect of the financial crisis on TFP. This set includes
information on the reduction of turnover, investment
and innovations. These are all expected to negatively
impact on TFP growth. A second set of dummies
(INTANG) captures firms decisions in relation to
invest in intangible assets and provides information
on whether the firm has conducted R&D investments
over the period91, employed a higher proportion of
educated workers compared to the national average
(human capital) or implemented some relevant
organisational changes. Extensive evidence at the
micro economic level shows
stock of the company, enhancing its productivity
performance and, more generally, the degree of
competitiveness (Hall et al. 2009, OMahony and
Vecchi 2009).
Another set of controls looks for firms engagement
in the international market (OPENESS). These
include firms decision to import material or service
intermediate inputs in 2008 or before, and the
decision to carry out FDI. An additional dummy
variable for companies belonging to foreign groups is
also considered to assess whether there is a positive
relationship between firms participation in
international networks and productivity growth. The
existing literature generally supports the evidence that
international firms are more productive than those
less prone to undertake foreign activities (Wagner
2012). However, little is known about the
performance of these firms during particularly critical
economic conditions.
The model also accounts for the role of institutional
settings on firms productivity, consistently with the
analysis in previous sections of the chapter. The
information on the EFIGE data set makes it possible
to check how companies adjusted their activities as a
result of the changes in labour market regulations
throughout the 2000s (LABOUR variable in the
econometric model 3.5). The impact of such reforms
is captured by a dummy variable which identifies
companies resorting to temporary and part-time
contracts. Existing evidence shows that a high share
of temporary workers is negatively associated with
productivity growth, due to the low experience and
low endowments of firm-specific human capital of
such employees (Daveri and Parisi 2010).
that

90
The cross-sectional regression model is estimated by OLS,
using heteroskedasticity robust standard errors.
91
R&D-performing firms is the preferred measure of innovation
effort among those available as its effect is more robust across
specifications.

90
Table 3.12. Determinants of TFP growth 2008-2009: OLS regression (by total sample, size classes and Pavitt groups)
Dependent variable: TFP change 2008-2009
(1) (2) (3) (4) (5) (6) (7) (8)
Total Small Medium Scale High- Specia-
Large firms Traditional
sample firms firms intensive tech lized
Log TFP 2008 0.881*** 0.880*** 0.903*** 0.843*** 0.898*** 0.900*** 0.886*** 0.868***
(0.010) (0.012) (0.017) (0.043) (0.017) (0.054) (0.019) (0.016)
Crisis effect
Turnover reduction -0.063*** -0.065*** -0.048*** -0.079*** -0.063*** -0.074*** -0.055*** -0.063***
(0.005) (0.005) (0.010) (0.021) (0.010) (0.021) (0.011) (0.006)
Investment reduction -0.019*** -0.013** -0.034*** -0.025 -0.019** -0.013 -0.032*** -0.013**
(0.004) (0.005) (0.009) (0.019) (0.009) (0.026) (0.011) (0.006)
Innovation reduction -0.006 -0.008 -0.003 0.018 -0.018** 0.030 0.008 -0.007
(0.004) (0.005) (0.008) (0.018) (0.009) (0.023) (0.010) (0.006)
Intangibles
R&D-doing firm 0.011** 0.011** 0.009 0.014 0.031*** -0.010 -0.004 0.005
(0.005) (0.005) (0.009) (0.021) (0.010) (0.024) (0.010) (0.006)
Human capital 0.008* 0.011** -0.006 0.010 0.006 -0.034 -0.002 0.015**
(0.005) (0.006) (0.010) (0.025) (0.010) (0.021) (0.010) (0.007)
Organizational change 0.003 -0.002 0.003 0.019 0.003 -0.016 -0.007 0.007
(0.004) (0.006) (0.009) (0.017) (0.009) (0.026) (0.009) (0.006)
Controls
Openness
Importer of materials 0.005 0.009 -0.009 0.005 0.003 0.030 0.010 0.002
(0.004) (0.005) (0.009) (0.020) (0.010) (0.023) (0.010) (0.006)
Importer of services 0.013** 0.011 0.013 0.020 0.017 0.047** -0.001 0.012
(0.005) (0.007) (0.009) (0.019) (0.011) (0.023) (0.010) (0.007)
FDI active 0.012 0.009 0.019 0.010 0.056** 0.006 0.003 -0.013
(0.011) (0.031) (0.015) (0.019) (0.024) (0.033) (0.019) (0.018)
FDI passive (foreign group) 0.022*** 0.032** 0.012 0.045** 0.017 0.037 0.014 0.029*
(0.008) (0.016) (0.011) (0.020) (0.014) (0.033) (0.015) (0.016)
Labour input
Flexible contracts -0.003 0.002 -0.036** 0.018 0.006 0.001 -0.037*** 0.002
(0.006) (0.007) (0.017) (0.028) (0.014) (0.034) (0.012) (0.008)
Financial input
Rationed credit -0.017 -0.027* 0.026 0.011 0.010 -0.028 -0.041 -0.020
(0.012) (0.014) (0.018) (0.051) (0.023) (0.045) (0.032) (0.015)
Other firm characteristics
Family management -0.013*** -0.012** -0.023 0.052 -0.004 -0.007 -0.031** -0.013**
(0.005) (0.005) (0.020) (0.040) (0.011) (0.030) (0.014) (0.006)
Quality certification 0.009** 0.009* 0.007 0.007 -0.005 0.032 0.032*** 0.009
(0.004) (0.005) (0.009) (0.021) (0.009) (0.028) (0.010) (0.006)
Young (less than 6 yrs) -0.001 0.008 -0.026*** -0.032 -0.008 -0.057** 0.014 0.000
(0.004) (0.005) (0.009) (0.025) (0.009) (0.027) (0.009) (0.006)

Constant 0.081*** -0.005 0.083** 0.150*** 0.055* 0.123 0.126*** 0.046*

(0.021) (0.043) (0.036) (0.047) (0.031) (0.082) (0.031) (0.025)

Size dummies Yes No No No Yes Yes Yes Yes

Pavitt dummies Yes Yes Yes Yes No No No No
Country dummies Yes Yes Yes Yes Yes Yes Yes Yes

Observations 7,077 4,852 1,641 584 1,844 273 1,349 3,611

R-squared 0.878 0.844 0.882 0.867 0.883 0.900 0.872 0.865
Notes: Robust Standard errors in parentheses. ***,**,* significant at 1,5, and 10%.

91
Due to the financial nature of the downturn, it is also
important to check whether worsening conditions in
the credit market affected firms performance
(variable FINANCE in the econometric model 3.5).
This is captured by the inclusion of a dummy variable
that takes value of 1 if the firm required credit during
the crisis but did not obtain it. There is considerable
evidence that EU SMEs were severely hampered in
accessing credit but the issue of whether this
translated into lower rates of productivity growth is
less explored (Houlton et al. 2012). Firms decisions
to invest in intangible assets and to compete on
international markets are related to their managerial
abilities (Castellani and Giovannetti 2010). The
analysis accounts for this issue by including a family
management variable (a dummy equals 1 if the share
of managers of the controlling family is higher than
the national mean) and a quality certification dummy,
which equals 1 if the firm has received some quality
equivalent certification. Managerial practices explain
large variation in firm productivity growth, negatively
affecting aggregate productivity growth (Bloom and
Van Reenen 2007, 2010).
Other control variables include the age of the firm,
country, size and industry (Pavitt) dummies. The
productivity impact of firms age is captured by a
dummy variable taking the unit value if the firm is
less than six year old. This controls for possible
differences in performance between young and
relatively old firms. The former have higher growth
potential, but they are not necessarily as productive
(or efficient) as the incumbents. During the 1990s,
highly innovative start-ups were found to make an
important contribution to aggregate productivity
growth in the US, whilst their role in the EU was less
clear (Bassanini and Scarpetta 2002).
Regression results are presented in Table 3.12.
Estimates are provided for the overall sample, and
then split according to size classes and Pavitt groups.
Overall, results corroborate most of the trends
highlighted in the earlier sections. The positive
coefficient on the level of TFP in 2008 suggests that
there has been no productivity catch-up among
companies during the crisis. This feature is common
across size classes and industry groups. This implies
that, within any single country, those firms which
were most productive at the outset of the downturn
accommodated better the negative shock, with above-
average outperformance in terms of TFP growth
during 2008-2009.
Firms which experienced a turnover and an
investment reduction underperformed compared to
those less affected by the crisis. The significance of
both variables indicates that the downturn impacted
distinctly on these two dimensions of firm
performance and this, in turn, negatively affected TFP
growth. The effect of turnover reduction was
relatively more severe among large and high-tech
92
firms, while the decline in investment particularly
affected productivity growth in medium firms and
specialised producers. On average, firms endowed
with intangible assets performed better, especially
when these factors were the outcome of research
activities or resulted from the employment of highly
educated workers. Results in Table 3.12 indicate that
intangible assets are important drivers of productivity
growth, particularly for small firms and those
specialised in scale-intensive production. The fact
that intangible inputs were not associated with better
productivity performance in more technologically
advanced productions, such as science-based
industries, may be due to the low variation among
these firms in R&D engagement and human capital
endowment. Interestingly, companies that had
undertaken organisational changes before the crisis
did not display a different productivity performance
from the rest of the sample during the downturn.
This may reflect the long time necessary before these
changes affect productivity (Rincon et al. 2012).
International firms were more productive with respect
to those active solely in the domestic market, in
particular those affiliated to a foreign group. This
feature is common to both small and large firms.
Conversely, the industry breakdown does not offer
insights on the role of this variable among Pavitt
categories, apart from the weakly significant effect
found for traditional firms. Among high-tech firms,
productivity grew faster in those importing service
intermediate inputs; scale intensive firms that had
previously carried out some production tasks abroad
also experienced higher productivity growth (FDI
active). In accordance with the literature on the
regulation governing the labour market, firms relying
upon flexible contracts experienced lower
productivity growth.However, this effect is confined
to medium-sized firms and specialised suppliers.
Results also provide evidence on the negative impact
of worsening credit conditions on productivity for
small firms (Houlton et al. 2012). Concerning other
firm characteristics, results show that family
management is another condition which hampered the
productivity growth of smaller firms, traditional
companies and specialised suppliers. Apart from this
last group of firms, going through quality certification
does not signal better managerial practices and, as a
consequence, faster productivity growth. Looking at
the age profile, productivity did not grow at a
differential speed among young firms; rather, young
medium-sized and high-tech firms underperformed
compared to more mature firms. Probably, young
companies were not sufficiently structured to tackle
the collapse of the market between 2008 and 2009, or
were not able to fully exploit their growth potential
due to the worsening in demand conditions.
3.5. SUMMARY AND POLICY IMPLICATIONS
This chapter has described the main features of recent
EU aggregate and industry productivity performance
from an international perspective, and has provided
empirical evidences on the key forces behind the
productivity growth differentials with respect to the
US. Identifying these factors is crucial for designing
policies able to reduce the gap and to restore
sustained growth in Europe. In the late 1990s,
characterised by the emergence of ICT technologies,
market services were the main culprit of the EU
productivity disadvantage. In the years leading to the
financial crisis, however, the EU experienced strong
ICT-related labour productivity growth in these
sectors, mirroring earlier developments in the US, and
boosting convergence towards US productivity levels.
Since the crisis, however, the EU-US productivity
gap has widened again.
The responses to the downturn have been
heterogeneous across the different sectors in the EU.
Overall, service sectors appear to have been relatively
less affected by the global economic crisis compared
to manufacturing. Some key sectors, such as business
services, have helped in narrowing the productivity
gap with the US in recent years. A plausible
explanation is that these sectors are usually more
sheltered from international competition than
manufacturing sectors, and therefore, less exposed to
global economic shocks. It is also possible that these
sectors have continued to reap the benefits of strong
tangible and intangible investments undertaken over
the past decade.
This positive outlook for labour productivity
expansion in some EU services sectors, however,
contrasts with a poor TFP performance. This finding
suggests that the EU continues to experience lower
levels of efficiency, with which inputs are used in the
production process, than the US. These results call for
a more in-depth analysis of factors affecting TFP
since, this is one of the main engines for increasing
income levels in the long run.
The econometric analysis has considered two main
channels through which it is possible to raise the
productivity growth potential and close the gap with
technology leaders. First, consideration has been
given to the role of absorptive capacity and
knowledge-base (intangible) assets i.e. R&D and
human capital in activating, and benefiting from,
international technology transfers. This mechanism
has been found in the literature to be highly
conducive to productivity growth through spillovers.
However, its growth-enhancing effect is
heterogeneous and it requires the ability to
accommodate the inflow of new technological
knowledge by re-allocating factors or, for instance,
expanding new product lines. The set of rules that
regulate the functioning of internal markets is
93
important in process of checking for the productivity
effects of technology transfers. This analysis has not
found convincing evidence that more restrictive
regulations on the labour, product and financial
markets significantly hamper the capacity of a
country to reap the benefits of knowledge developed
elsewhere.
The second channel, is the role of production
efficiency as a possible driving force behind the
widening productivity gaps between the EU and the
US. The analysis has shown that productive
efficiency is significantly higher in countries with less
restrictive product market regulations or employment
protection laws. However, when there are few
restrictions on the use of temporary contracts and in
financial markets, the efficiency gaps with respect to
the frontier are likely to increase, as these might
encourage firms to adopt cost-cutting strategies rather
than the most efficient methods of production.
Investment in ICT assets, on the other hand, is one of
the crucial factors that help in reducing the distance
from the most efficient country and/or industry.
Broadly consistent evidence also emerges from the
analysis of firm-level performance at the outset of the
financial crisis. The analysis undertaken for seven EU
economies provides strong micro foundations to the
observed widening productivity differentials at a
more aggregate level. This study shows that the most
productive firms, prior to the crisis, which
experienced faster TFP growth afterwards. This
finding confirms that, even within the EU, the recent
downturn seems to have reinforced the trend of
diverging productivity patterns which emerged in the
earlier period.
Overall, the analysis carried out in this chapter
provides insights into which policies may be more
effective in raising productivity performance within
the EU and closing the gap with the US. A common
finding throughout the chapter is that intangible assets
(R&D, human capital, organizational change) are
important sources of TFP growth and sustained long-
run competitiveness. From this perspective,
initiatives aimed at stimulating such investments may
be particularly useful.
Albeit EU countries represent an important share of
R&D at a worldwide level, they are less specialized
in high-tech sectors compared to the US. The
different structural specialisation of the EU countries
explains why they have fewer young firms among its
leading innovators and their young firms are less
innovative than in the US (Cincera and Veugelers
2010). Also, in Europe the proportion of R&D-doing
firms is considerably lower than the US. These
factors can explain the discussed research gap
(Moncada-Patern-Castello et al. 2010).
Ample evidence can be found in the literature, for
example, about the effectiveness of tax incentives to
raise research effort. This policy instrument does not
distort market incentives because; it reduces the cost
of R&D without influencing firms choices regarding
specific projects (David et al. 2000)92. In the OECD
area, public policies to directly sustain R&D have
also been implemented through a combination of
measures favouring a large spectrum of knowledge-
intensive sectors (ICT, pharmaceuticals,
biotechnologies, etc.). In this way, these countries
have sought to shift their industrial structures towards
high-tech productions and increase thus their
international competitiveness.
Specific policy initiatives may also be put in place to
increase a firms endowment of qualified workers, for
instance facilitating hiring of highly qualified workers
(such as professional managers) or to sustain
workforce training to enhance the endowment of
firm-specific human capital. Other sound policies
could be directed towards raising investment in inputs
such as ICT, which can assist in the reorganisation of
production. Specific ICT applications, such as
enterprise software systems, have been related to
increasing productivity at the firm level in the
existing literature (Engelstatter 2009). These
measures would also be viable for smaller firms
which do not always have the necessary resources to
embark on formal R&D activities and need
92
A possible drawback is that it influences the composition of
research project favouring those with a higher profitability in
the short run.
94
alternative ways of increasing their competitiveness
(EC 2012 and 2013). It should be borne in mind that
ICT may spur productivity performance by increasing
efficiency in production tasks and this effect may
occur with some lags. These measures may therefore
be accompanied by other policies targeted to
facilitating factors. Policies aimed at improving the
functioning of product and factor markets may be
particularly effective. Reducing the strictness of
product market regulations, largely concentrated in
key service-providing industries, is likely to be
conducive to higher levels of efficiency across the
whole economy, by allowing input re-allocation,
outsourcing of marginal tasks, and the adoption of the
best production and managerial practices. Changes in
the regulatory setting of the labour market should also
be tailored to restore an optimal mix of regular and
temporary workers, bearing in mind that an excessive
liberalisation of temporary workers contracts may
hinder productivity and efficiency performance.
Given the role of financial input on productivity
performance, it appears useful to promote policies
designed to increase firms access to external
funding, such as bank credit and private bonds. These
measures should be conceived and applied within an
appropriate regulatory framework which will
safeguard the stability and facilitate the reduction of
productivity and efficiency gaps.
 REFERENCES
Acharya, R., and Basu, S., (2010), ICT and TFP Growth: Intangible Capital or Productive Externalities?,
Industry Canada Working Paper 2010-1.
Aghion, P., and Howitt, P., (1992), A model of growth through creative destruction, Econometrica, 60(2), pp.
323-351.
Aghion, P., (2002), Schumpeterian growth theory and the dynamics of income inequality, Econometrica, 70(3),
pp. 855-882.
Altomonte, C., and Aquilante, T., (2012), The EU-EFIGE/Bruegel-Unicredit Dataset, Bruegel Working Paper
2012/13.
Ang, J.B., (2011), Financial development, liberalization and technological deepening?, European Economic
Review, 55, pp. 688-701.
Ang, J.B., and Madsen, J.B., (2012), Risk capital, private credit and innovative production, Canadian Journal
of Economics, 45(4), pp. 1608-1639.
Arnold, J. Nicoletti G., and Scarpetta S., (2011), Does anti-competitive regulation matter for productivity?
Evidence from European firms, IZA Discussion Papers 5511.
Bassanini A., Nunziata, L., and Venn, D., (2009), Job protection legislation and productivity growth in OECD
countries, Economic Policy, 24, pp. 349-402.
Bassanini, A, Scarpetta, S., (2002). Growth, Technological Change, and ICT Diffusion: Recent Evidence from
OECD Countries, Oxford Review of Economic Policy, 18(3), pp. 324-344.
Basu, S., Fernald, J.H., Oulton, N., and Srinivasan, S., (2004), The case of the missing productivity growth, or
Does information technology explain why productivity accelerated in the United States but not in the United
Kingdom?, in Gertler, M., and K. Rogoff, (Eds), NBER Macroeconomics Annual 2003.
Beck, T., Demirg-Kunt, A., Levine, R. (2009), Financial Institutions and Markets Across Countries and Over
Time: Data and Analysis, World Bank Policy Research Working Paper n. 4943.
Bloom, N., Schankerman, M., and van Reenen, J. (2013), Identifying Technology Spillovers and Product
Market Rivalry, Econometrica, forthcoming.
Bloom, N., van Reenen, J., (2007), Measuring and Explaining Management Practices Across Firms and
Nations Quarterly Journal of Economics, 122 (4), 1351-1408
Bloom, N., van Reenen, J., (2010), Why do management practices differ across firms and countries?, Journal
of Economic Perspectives, 24 (1), pp.203-224.
Bourls R., Cette G., Lopez J., Mairesse J., and Nicoletti G., (2012), Do product market regulations in upstream
sectors curb productivity growth? Panel data evidence for OECD countries, Review of Economics and
Statistics, forthcoming.
Breitwieser, A., and Foster, N., (2012), Intellectual property rights, innovation and technology transfer, WIIW
Working Paper no. 88, Vienna Institute for International Economic Studies.
Brynjolfsson, E., Hitt, L. , (2000), Beyond computation: information technology, organizational transformation
and business performance, Journal of Economic Perspectives, 14, pp.2348.

Brynjolfsson, E., Saunders, A.,(2009), Wired for Innovation: how information technology is reshaping the
economy. MIT Press.
Buccirossi, P., Lorenzo Ciari, L., Duso, T., Spagnolo, G., and Vitale, C., (2012), Competition policy and
productivity growth: An empirical assessment, Review of Economics and Statistics, forthcoming.
Card, D., and Krueger, A.B., (1995), Time-series minimum-wage studies: A meta-analysis, American
Economic Review, 85(2), pp. 238-243.
Castellani, C.and Giovannetti, G., (2010), Productivity and the international firm: dissecting heterogeneity,
Journal of Economic Policy Reform, 13(1), pp. 25-42.
Cincera, M. and Veugelers, R. (2010), Young leading innovators and EUs R&D intensity gap, JRC-IPTS
Working Papers on Corporate R&D and Innovation, 11/2010.

95
Cohen, W., and Levinthal, D., (1989), Innovation and learning: The two faces of R&D, Economic Journal, 99,
pp. 569-596.
Conway, P., De Rosa, D., Nicoletti, G., and Steiner, F., (2006), Regulation, competition and productivity
convergence, OECD Economic Studies, 2, 9.
Crespo-Cuaresma, J., Foster, N., and Scharler, J., (2004), On the determinants of absorptive capacity: Evidence
from OECD Countries, Current Issues in Economic Growth, Proceedings of OeNB Workshops, Vienna,
Austrian National Bank, pp. 58-81.
Currie, J. and Fallick, B.C. (1996), The minimum wage and the employment of youth: Evidence from the
NLSY, Journal of Human Resources, 31(2), pp. 404-428.
Damiani, M., Pompei, F., and Ricci, A., (2013), Temporary job protection and productivity growth in EU
economies, MPRA Paper 29698 2011. Revision January 2013.
Darby, J., Hart, R.A., Vecchi, M., (2001), Wages, work intensity and unemployment in Japan, UK and USA.
Labour Economics 8, pp. 243-258.
Daveri, F. Parisi, M.L., (2010), Experience, Innovation and Productivity Empirical evidence from Italy's
slowdown, CESifo Working Paper Series 3123, Munich, CESifo Group.
David, P.A., Hall, B.H., Toole, A.A., (2000), Is public R&D a complement or substitute for private R&D? A
review of the econometric evidence, Research Policy 29, pp. 497529.
Dedrick, J., Gurbaxani, V., and Kraemer, K.L., (2003), Information technology and economic performance: a
critical review of the empirical evidence, ACM Computing Surveys 35 (1), 1-28.
Dietzenbacher, E., Los, B., Stehrer, R., Timmer, M. and deVries, G., (2013), The Construction of World Input-
Output Tables in the WIOD project, Economic Systems Research, 25(1), pp. 71-98.
Eaton, J. and S. Kortum, (1999), International technology diffusion: Theory and measurement, International
Economic Review, 40, pp. 537-570.
EC, (2008), Sectoral Growth Drivers and Competitiveness in the European Union, Michael Peneder (Ed.),
European Commission, DG Enterprise and Industry.
EC (2012), Annual growth survey 2013, European Commission, Brussels.
EC (2013), Supporting Small- and Medium-sized enterprise in 2012, A joint report of the European
Commission and the European Investment Bank Group, Brussels.
Engelsttter, B. (2009), Enterprise Systems and Labour Productivity: Disentangling Combination Effects. ZEW
Working Papers n. 09-040.
Evenson, R., and Westphal, L., (1995), Technological change and technology strategy In J. Behrman and T.
Srinivasan (eds.), Handbook of Development Economics, vol. 3A, pp. 2209-2229, Amsterdam: North Holland
Publishing Company .
Falvey, R.E., Foster, N. and Greenaway, D., (2007), Relative backwardness, absorptive capacity and knowledge
spillovers, Economics Letters 97, pp. 230-234.
Farrell, M.J., (1957), The Measurement of Productivity Efficiency,Royal Statistical Society, 120, pp. 253-290.
Fiori, G., Nicoletti G., Scarpetta, S., and Schiantarelli, F., (2012), Employment effects of product and labour
market reforms: Are there synergies?, Economic Journal, 122, pp. F79-F104.
Ginarte, J.C, and Park, W.G, (1997), Determinants of patent rights: A cross-national study, Research Policy,
26(3), pp. 283-301.
Goodridge, P., Haskel, J. And G. Wallis (2013), Can Intangible Investments explain the UK Productivity
Puzzle?, National Institute Economic Review 224, May 2013.
Gordon, R.J., (2011), The evolution of Okuns law and of cyclical productivity fluctuations in the United States
and in the EU-15, paper presented at ES/IAB Workshop, Labor Market Institutions and the Macroeconomy,
Nuremberg, June 17-18, 2011.
Griffith R., Redding S., and van Reenen J., (2004), Mapping the Two Faces of R&D: Productivity Growth in a
Panel of OECD Industries, Review of Economics and Statistics 86, 883895.
Griffith, R., Harrison, R., and Macartney, G., (2007), Product market reforms, labour market institutions and
unemployment, Economic Journal 117, pp. C142-C166.

96
Griliches, Z. and Mairesse, J. (1984), Productivity and R&D at the Firm Level, in R&D, Patents and
Productivity Griliches, Z. (Ed.), Chicago: University of Chicago Press.
Guellec, D., and van Pottelsberghe de La Potterie, B., (2004), From R&D to Productivity Growth: Do the
Institutional Settings and the Source of Funds of R&D Matter?, Oxford Bulletin of Economics and Statistics,
66:3, pp. 353-378.
Hall, B.H., F. Lotti, and Mairesse, J., (2009), Innovation, R&D and Productivity in SMEs: Evidence for Italy,
Small Business Economics, 33(1), pp. 13-33.
Hansen, B.E., (1996), Inference when a nuisance parameter is not identified under the null hypothesis.
Econometrica, 64, 413-430.
Hansen, B.E., (1999), Threshold effects in non-dynamic panels: Estimation, testing and inference. Journal of
Econometrics, 93, 345-38.
Hansen, B.E., (2000), Sample splitting and threshold estimation. Econometrica, 68, 575-603.
Hornsteien, A., and Krusell, P., (1996), Can Technology Improvements Cause Productivity Slowdowns?, in
Macroeconomics Annual 1996, Bernanke, B. S., and Rotemberg, J.J. (Eds.), Cambridge: MIT Press.
Houlton, S., Lawless, M., McGunn, F. (2012), Firm credit in Europe: A tale of three crises Central Bank of
Ireland, Research Technical Paper 4.
Inklaar, R. and Timmer, M.P. (2008), GGDC Productivity Level Database: International Comparisons of
Output, Inputs and Productivity at the Industry Level, Groningen Growth and Development Centre Research
Memorandum GD-104, Groningen: University of Groningen, September 2008".
Jorgenson, D.W., and Nomura, K., (2007), The industry origins of the U.S.-Japan productivity gap, Economic
Systems Research, 19(3), pp. 315-412.
Kang, L., O'Mahony, M., and Peng, F., (2012), New measures of workforce skills in the EU National Institute
Economic Review, 220(1), pp. R17-R28.
Kretschmer, T., (2012), Information and communication technologies and productivity growth: A survey of the
literature, OECD Digital Economy Papers, No. 195.
Kumbhakar, S., Ortega-Argils, R., Potters, L., Vivarelli, M. and Voigt, P.,(2010), Corporate R&D and firm
efficiency: Evidence from Europes top R&D investors, European Commission, Joint Research Centre,
24325 EN/11.
Lane, P.R., and Milesi-Ferretti, G.M, (2010), The cross-country incidence of the global crisis, CEPR
Discussion Papers 7954.
Mas, M. (2010), Productivity in the advanced countries: from expansion to crisis, in Industrial Productivity in
Europe: Growth and Crisis, Mas, M. and Stehrer, R. (eds), Edward Elgar, UK.
Maskus, K.E., Neumann, R., and Seidel, T., (2012), How national and inter-national financial development
affect industrial R&D, European Economic Review, 56, pp. 72-83.
Mason, G., OLeary, B., and Vecchi, M., (2012), Certified and uncertified skills and productivity growth
performance: Cross-country evidence at industry level, Labour Economics, 19(3), pp. 351-360.
Moncado-Patern-Castello, Ciupagea, C., Smith, P., Tbke,K.and Tubbs, M. (2010), Does Europe perform too
little corporate R&D? A comparison of EU and non-EU corporate R&D performance, Research Policy 39,
pp. 523-536.
Nicodeme, G., and Sauner-Leroy, J.B., (2004), Product market reforms and productivity: a review of the
theoretical and empirical literature on the transmission channels, European Economy - Economic Papers
218, DG ECFIN, European Commission.
Neumark, D., and Wascher W.L., (2008), Minimum Wages, The MIT Press.
Nishioka, S. and Ripoll, M., (2012), Productivity, trade and the R&D content of intermediate inputs, European
Economic Review, 56, pp. 1573-1592.
OECD (2011), Economic Survey of Germany 2010: Structural reforms to lift potential growth in a globalised
world, OECD, Paris.
OECD (2012), OECD Compendiumof Productivity Indicators 2012, OECD, Paris.

97
OMahony, M. and van Ark, B., (2003), EU productivity and competitiveness: An industry perspective. Can
Europe resume the catching-up process?, European Commission, Enterprise publications.
OMahony, M. and Timmer, M.P., (2009), Output, input and productivity measures at the industry level: the EU
KLEMS database, Economic Journal, 119(538), pp. F374-F403.
O'Mahony, M., and Vecchi, M., (2009), R&D, knowledge spillovers and company productivity performance,
Research Policy, 38(1), pp. 35-44.
Parente. S.L., and Prescott, E., (1994), Barriers to technology adoption and development, Journal of Political
Economy, 102, pp. 298-321.
Parente. S.L., and Prescott, E. , (1999), Monopoly rights: A barrier to riches, American Economic Review, 89,
pp. 1216-1233.
Parente. S.L., and Prescott, E. , (2003), Barriers to Riches. Cambridge: MIT Press.
Park, W.G., (2008), International patent protection: 1960-2005, Research Policy, 37, pp. 761766.
Pavitt, K., (1984), Sectoral patterns of technical change: Towards a taxonomy and a theory, Research Policy,
13, pp. 343-373.
Rajan, R.G., and Zingales, L., (1998), Financial dependence and growth, American Economic Review, 88(3),
pp. 559-586.
Romer, P., (1990), Endogenous technological change, Journal of Political Economy, 98(5), pp. S71-S102.
Rincon, A. Vecchi, M., Venturini, F., (2012), ICT spillovers, absorptivecapacity and productivity performance,
Quaderni del Dipartimento di Economia, Finanza e Statistica 103/2012, Universit di Perugia, Dipartimento
Economia, Finanza e Statistica.
Schankerman, M., (1981), The effect of double-counting and expensing on the measured returns to R&D,
Review of Economics and Statistics, 63(3), pp. 453-458.
Stiroh, K.J., (2002), Are ICT spillovers driving the New Economy?, Review of Income and Wealth, 48(1), pp.
33-57.
Timmer, M., Inklaar, R., OMahony, M., and van Ark, B., (2010), Economic Growth in Europe, Cambridge
University Press, Cambridge, UK.
Timmer, M., van Moergastel, T., Estuivenwold, E., Ypma. G.,O'Mahony, M. and Kangasniemi, M. (2007),
EUKLEMS Growth and Productivity Account: Methodology Part 1, EU KLEMS project.
UNIDO (2002), Industrial Development Report 2002/2003. Vienna, United Nations Industrial Development
Organization.
Tbke, A., Hervs, F. and Zimmermann, J.: The 2012 EU Survey on R&D Investment Business Trends,
European Commission, Joint Research Centre, JRC72991, EUR 25424 EN, p.21.
Van Ark, B., Levanon, G, Chen, V. and Cheng, B., (2012), Performance 2011: Productivity, employment and
growth in the worlds economies, Research Report R_-1475-11-RR, The Conference Board.
Van, Ark ,B. (2008), O'Mahony, M. and M.P.Timmer (2008), The Productivity gap between Europe and the
US: Trends and causes, Journal of Economic Perspectives, volume 22, number 1, pp 25-44.
Venturini, F. (2009), The long-run impact of ICT, Empirical Economics, 37(3), pp. 497-515.
Visser, J. (2011), ICTWSS: Database on institutional characteristics of trade unions, wage setting, state
intervention and social pacts in 34 countries between1960 and 2007, Amsterdam Institute for Advanced
Labour Studies (AIAS), University of Amsterdam.
Wagner, J., (2012), International trade and firm performance: a survey of empirical studies since 2006, Review
of World Economics (Weltwirtschaftliches Archiv) 148, pp. 235-267.
Wlfl, A., Wanner, I., Kozluk, T., and Nicoletti, G., (2009), Ten years of product market reform in OECD
countries Insights from a revised PMR indicator, OECD Working Papers no. 695, OECD,Paris.
World Bank (2012), Doing Business 2012: Doing Business in a More Transparent World, World Bank.
Yanikkaya, H., (2003), Trade openness and economic growth: A cross-country empirical investigation, Journal
of Development Economics, 72(1), pp. 57-89.

98
 ANNEX 1
ABBREVIATIONS
ECR Enforcing Contract Time
EPC Employment Protection Legislation for collective dismissals.
EPL Overall Employment Protection Legislation
EPR Employment Protection Legislation for regular contracts
EPT Employment Protection Legislation for temporary contracts
EU-15 Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg,
Netherlands, Portugal, Spain, Sweden, United Kingdom.
EU-27 Austria, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia, Finland, France,
Germany, Greece, Hungary, Ireland, Italy, Lithuania, Luxembourg, Latvia, Malta, Netherlands,
Poland, Portugal, Romania, Spain, Sweden, Slovenia, Slovakia, UK.
EU-8 Austria, Belgium, France, Netherlands, Spain, Germany, Italy, UK.
FDI Foreign Direct Investment
GDP Gross Domestic Product
GPT General Purpose Technology
ICT Information and Communications Technologies
IPR Intellectual Property Rights
JP Japan
LC Labour Composition
LP Labour Productivity
NACE National Classicisation of Economic Activities
NMW National Minimum Wage
NUTS The Nomenclature of Territorial Units for Statistics
OECD Organisation for Economic Cooperation and Development
OLS Ordinary Least Squares
PMR Product Market Regulation
R&D Research and Development
RI Regulation Impact
SFA Stochastic Frontier Analysis
SME Small and Medium-sized Enterprise
TE Technical Efficiency
TFP Total Factor Productivity
US United States

99

2.A. METHODOLOGY TO
AGGREGATE PRODUCTIVITY DATA
The Conference Board Total Economy Database
(TED) and the EUKLEMS database contain a wide
range of economic performance measures on a
country-by-country basis. This information is
however of limited use if one wants to establish
meaningful comparisons of growth and productivity
trends between the whole of the EU and other world
economies. In this study, the methodology set out in
Timmer et al, 2007 is followed to construct EU
aggregate (e.g. the EU-27 or EU-15) measures of
output, input and productivity.
This methodology is based in the use of a Trnqvist
quantity index, which is a discrete time
approximation to a Divisia index. A Divisia index
defined as a continuous-time weighted sum of the
growth rates of various components, where the
weights are the component's shares in total value; in
the Trnqvist index, the growth rates are defined as
the difference in the natural logarithm of consecutive
observations of the components, and the weights are
equal to the mean of the factor shares of the
components in the corresponding pair of time
periods (e.g. years).
Aggregation over countries
The derivation of an aggregate measure of labour
productivity for the EU-27 is outlined in practical
terms here. First of all, annual growth rates of labour
productivity for each of the EU-27 countries are
computed for the time period under consideration (as
the differences in the natural logarithms). Secondly,
the annual shares of each country in EU nominal
output are calculated using PPP-converted values,
which adjust for purchasing power parities price
differentials across countries 93 (Inklaar and Timmer,
2008).
The calculation of nominal output shares for each of
the EU countries is given by the following
expression:
[ ]
93
Purchasing Power Parities (PPP), which are available for
economy level and for detailed industries, are usually given
for a benchmark year. Here the PPPs are given for 1997 (See
Inklaar and Timmer, 2008).
100
ANNEX 2
METHODOLOGIES
where denotes PPP-adjusted nominal output;
denotes nominal output; denotes country, denotes
industry, and denotes year.
The overall labour productivity for the EU-27 is then
calculated as a weighted average of country
productivity growth rates, as set out below:
where denotes the two-year average shares of
each country in total nominal output. Once the
annual growth rates for the EU country grouping are
obtained, it is feasible to construct an aggregate
index of labour productivity in relation to a base
year (for example, assuming that labour productivity
is equal to 100 in year 1995).
Aggregation over industries
A similar procedure to the one outlined above can be
applied to calculate aggregate performance for a
specific group of industries. For instance, to measure
productivity growth in the high-technology
manufacturing sector of a particular country.
Moreover, if productivity in the high-technology
manufacturing sector of the EU as a whole wants to
be computed, a double aggregation procedure has to
be followed. First an aggregation is performed over
countries, and then, over industries, following
recommendations in Timmer et al (2007).
2.B. GROWTH ACCOUNTING
METHODOLOGY
The growth accounting methodology (Jorgenson and
Griliches, 1967; Jorgenson et al, 1987) has been
widely used to assess the contribution of the
different factors of production to aggregate
economic growth. According to this methodology,
which is rooted in neoclassical theory, the part of
output growth that is not accounted by the growth in
inputs, usually capital and labour can be attributed to
TFP, a proxy measure for technological progress.
Assuming a Cobb-Douglas production function,
output (i.e. value added) is a function of capital
( , labour ( and technology ( in the following
terms:

Assuming that factor markets are competitive, full
input utilisation and constant returns to scale, the
growth of output can be expressed as the cost-share
weighted growth of inputs and technological change
(A), using the translog functional form common in
such analyses:
where and are the two-period average share of
labour and capital input in nominal output.
With the use of this empirical approach it is possible
to identify and quantify the role of labour and capital
in aggregate growth. More recent contributions
extended the framework to allow for the separate
analysis of ICT assets (computers, software and
communications) and non-ICT capital assets
(machinery, transport equipment, residential
buildings, infrastructure), as well as for changes in
workforce composition, in terms of labour
characteristics such as educational attainment, age or
gender (Jorgenson et al, 2005). Growth in output can
be decomposed into the following elements:
squares. This involves finding the value of
where the contribution of each factor input is given
by the product of its share in total costs and its
growth rate;
is the two-period average share of ICT assets in
total capital compensation; and is the two-
period average share of non-ICT assets in total
capital compensation.
The growth in labour input can be split into growth
of hours worked and changes in labour composition.
Labour composition in EUKLEMS is derived by
dividing labour into types and multiplying growth in
each type by wage bill shares.
. of the likelihood ratio test allowing one to obtain a
To analyse productivity it is useful to divide output
and inputs by the number of hours. The following
expression can be derived for labour productivity
growth:
( )
101
Based on the above formulae, the EUKLEMS and
The Conference Board Total Economy Database
provide a full decomposition of output and labour
productivity growth into the contributions of the
various factor inputs and TFP growth.
2.C. THRESHOLD REGRESSIONS
Threshold models have in recent times received a
great deal of attention as a means of modelling
parameter heterogeneity and non-linearities. In a
series of papers Hansen (1996, 1999 and 2000)
develops a technique that allows the sample data to
jointly determine both the regression coefficients
and the threshold value for OLS and (non-dynamic)
fixed effects panel models.
The threshold model for a single threshold can be
written as:
where is the indicator function and is the
threshold variable. Here the observations are divided
into two regimes depending on whether the
threshold variable is smaller or larger than . The
two regimes are distinguished by different
regression slopes, and . Chan (1993) and
Hansen (1999) recommend estimation of by least
that
minimises the concentrated sum of squared errors. In
practice this involves searching over distinct values
of for the value of at which the sum of squared
errors is smallest, which is then our estimate of the
threshold. Once we have an estimate for the
threshold it is straightforward to estimate the model.
Hansen (2000) extends this method to the case of
non-dynamic fixed-effects panel models.
Having found a threshold it is important to
determine whether it is statistically significant or
not, that is, to test the null hypothesis; .
Given that the threshold is not identified under
the null, this test has a non-standard distribution and
critical values cannot be read off standard
distribution tables. Hansen (1996) suggests
bootstrapping to simulate the asymptotic distribution
p-value for this test. Firstly, one estimates the model
under the null (i.e. linearity) and alternative (i.e.
threshold occurring at ). This allows one to
construct the actual value of the likelihood ratio test
:
where
Here and are the residual sum of squares from
the linear and threshold models respectively. Using a
parametric bootstrap (see Cameron and Trivedi,
2005) the model is then estimated under the null and
alternative and the likelihood ratio is calculated.
This process is repeated a large number of times.
The bootstrap estimate of the p-value for under
the null is given by the percentage of draws for
which the simulated statistic exceeds the actual
one.
The approach is also easily extended to consider
more than one threshold. While it is straightforward
to search for multiple thresholds, it can be
computationally time-consuming. Bai (1997) has
shown, however, that sequential estimation is
consistent, thus avoiding this computation problem.
In the case of a two-threshold model, this involves
fixing the first threshold and searching for a second
threshold. The estimate of the second threshold is
then asymptotically efficient, but not the first
threshold because it was estimated from a sum of
squared errors function that was contaminated by the
presence of a neglected regime. Bai (1997) suggests
estimating a refined estimator for the first threshold,
which involved re-estimating the first threshold,
assuming that the second threshold is fixed. The test
of significance of the second threshold proceeds
along the same lines as described above, with the
null and alternative hypotheses being of a one and
two threshold model respectively.
2.D. STOCHASTIC FRONTIER
ANALYSIS (SFA)
Frontier Analysis, initially developed in Farrell
(1957) and successively extended by Aigner et al.
(1977), Kumbhakar and Lovell (2000) and Greene
(2005) among others, aims to identify the production
frontier, i.e. the maximum level of output that can be
achieved by using the available inputs. Compared to
regression analysis, the estimation of the frontier
production function implies fitting a regression line
Y Y
* *
* * *
* *
* * * *
* *
X X
A: Regression analysis
102
over the units (industries in our case) that produce
the most output. The difference between the two
techniques can be easily seen in the following figure,
where Y indicates output and X denotes a generic
input:
Countries/industries at the frontier are those that are
making the most efficient use of their resources.
Those below the frontier have some level of
inefficiency, which can be directly estimated by the
distance between each industry and the frontier
industry.
It is possible to distinguish between two frontier
methods, Deterministic Frontier (DEA) and
Stochastic Frontier (SFA). DEA (Farrell 1957,
Charnes, Cooper and Rhodes 1978) provides a non-
parametric approach for estimating production
technologies and measuring inefficiencies in
production. It relies on the assumption that all
deviations from the frontier are caused by technical
inefficiency, without making allowanced for
measurement errors and/or random components.
This implies that not only the method is very
sensitive to the presence of outliers, but also lacks
the necessary diagnostic to help the user determine
whether or not the chosen model is appropriate,
which variables are significant and which are not.
These shortcomings are overcome by using the SFA.
Here the identification of the frontier technology is
based on the econometric estimation of a production
function, usually a Cobb-Douglas or a semi-translog
function. Differently from standard regression
analysis, frontier analysis allows for the presence of
a composite error term, which includes a random
component and an inefficiency term. The random
component allows for the presence of measurement
errors and other effects not captured by the model.
The inefficiency term measures technical
inefficiencies, i.e. the distance of each
country/industry from the frontier. This ranges
between 0 and 1, with higher values identifying
* *
* * *
* *
* *
* *
B: Frontier analysis
more efficient units. Technical efficiency scores
derived for each unit/industry can then be analysed
across different dimensions to pinpoint areas
characterised by low/high inefficiencies.
The performance of an industry depends not only on
the inputs used in the production process but also on
other external or environmental factors that can
affect the efficient use of resources. These are
usually factors that are outside the control of an
industry, even though it is possible that some factors
play a dual role, i.e. they affect both frontier output
and inefficiency (Kneller and Stevens 2006). The
SFA framework can easily account for this by
modelling the mean level of the inefficiency term as
a function of these additional factors. Production
frontier and determinants of inefficiency are
estimated simultaneously by maximum likelihood
(ML) (Battese and Coelli 1995).
2.E. PAVITT TAXONOMY (1984)
Using industry-specific characteristics of innovative
UK firms Pavitt (1984) identifies some major
technological trajectories in manufacturing, on the
basis of which it is possible to identify some specific
patterns of sectoral innovation. The Pavitt (1984)
taxonomy maps industries according to the source of
innovation activities made by the firms (internal vs
external), the nature of innovation (informal vs
formal, or learning vs R&D), firm size (small,
medium, or large), appropriability of innovation
(low vs high returns to innovation), method of
protection (secrecy vs patents) etc. Industries or
firms can be grouped into the following categories:
Scale-intensive: They are large firms exploiting
increasing returns to scale and learning-by-doing
associated with the size of the reference market, or
of their own plant. The source of innovation may be
both external and internal. In the former case, these
firms acquire production technologies from
specialised suppliers. In the latter case, in-house
R&D activities are performed to develop new types
of products; in this case, patenting is effective to
protect innovation. The main economic activities of
such firms are basic metals or the production of
durable goods.
103
Science based: They are mainly large firms using
internal sources of knowledge to produce
innovations (R&D). Their knowledge base is
complex and relies upon scientific advances.
Sometimes, innovations are developed between
private firms and universities and other research
institutes. Patents are the major, but not exclusive,
tools to protect innovations. Small firms may be
very competitive in certain technologically advanced
niches. The main economic activities of such firms
are pharmaceuticals, electronics, etc.
Specialised suppliers: They are small- and medium-
sized firms manufacturing sophisticated equipment
and/or precision machinery. They strongly rely upon
internal sources of innovation (engineering and
design capabilities are pivotal), developing new
products by continuously interacting with their
customers, i.e. downstream firms using in their
production the equipment developed by this
category. The nature of innovation of this type of
firms is therefore informal and based on learning.
Supplier dominated: They are traditional firms,
representing the least technologically advanced
branch of the manufacturing sector. Their main
source of innovation is external and consists in
introducing cost-saving process innovations, or
implementing advanced technologies, equipment
and materials, developed in other sectors. The only
internal source of innovation is the learning
associated with the usage of acquired inputs. Given
the low level of appropriability of internal
innovation, patenting is not very developed. The
main economic activities of such firms are food,
textile, footwear, etc.
104
 Chapter 4.
A MANUFACTURING IMPERATIVE IN THE EU: THE ROLE
The economic crisis changed the perceptions of the
role of the manufacturing sector in the economy.
Manufacturing has redeemed its reputation in the
sense that a comparatively large manufacturing sector
is no longer considered to reflect an outdated
economic structure, inadequate for a post-industrial,
services-dominated economy like the EU. Rather,
nurtured by the observation that within the EU,
countries which have maintained a larger
manufacturing base fared better during and after the
crisis (Reiner, 2012; Frst, 2013), a dynamic
manufacturing sector is again considered a
prerequisite for an innovative and fast-growing
economy. In a recent Communication, the European
Commission, emphasises that a vibrant and highly
competitive EU manufacturing sector is a key
element for solving societal changes ahead and a
more sustainable, inclusive and resource-efficient
economy (European Commission, 2010a).
This altered perception of manufacturing raised
concerns that manufacturing production had declined
too much (Warwick, 2013) in some Member States
leading to a loss of knowledge, capabilities and
supplier networks which have been referred to as the
manufacturing commons (Pisano and Shih, 2009)94.
Earlier arguments for a manufacturing imperative
(Rodrik, 2012) were re-discovered and the current
structural shift out of manufacturing in advanced
economies, including most EU Member States,
started to look less advantageous. The urge felt by
policy makers and the business community to
maintain a broad manufacturing base in Europe also
led to a renewed interest in industrial policy in
Europe and elsewhere (including the United States).
The importance of industrial structures is widely
accepted. The potential for economic policy to shape
that structure, however, remains highly disputed,
particularly in Europe where the track record of
interventionist industrial policy experiments in the
1960s and 1970s was rather disappointing (Crafts,
2010; Owen, 2012). Industrial policy, understood as
selective government interventions seeking to alter
the structure of production towards industries that are
94
The industrial commons are a reference to the commons which
is the land belonging to a (village) community as a whole and
which could also be used by each member of the community
(typically for grazing of animals). They can be described as the
general stock of knowledge, competences and skills (often
embodied in the workforce) and institutions (including supplier
networks) relevant for modern manufacturing activities that
can be shared and accessed by the manufacturing sector as a
whole (Pisano and Shih, 2009).
105
OF INDUSTRIAL POLICY
expected to offer higher growth prospects (Pack and
Saggi, 2006), can in principle try to foster structural
change towards any sector or industry that
government authorities consider to be strategic or
potential carrier of growth. Viewed through the lenses
of a manufacturing imperative, the particular
characteristics of manufacturing industries (such as
externalities and increasing returns to scale95) call for
industrial policies that re-direct the European
economy towards manufacturing activities and aim at
strengthening or restoring the industrial commons.
Despite this renewed debate about the objectives and
instruments of EU industrial policy, it remains deeply
rooted in the principles of competition, favouring
general framework policies (such as the proper
functioning of the Internal Market and competition
rules) and horizontal policies over sector-specific
interventions96. Nevertheless, in the aftermath of the
economic crisis the European Commissions focus on
framework policies has been supplemented with more
sector-specific policy objectives such as the definition
of key priority areas which include inter alia the
development of clean vehicles and vessels and smart
grids (European Commission, 2012a). Sector-specific
action may indeed be warranted in cases where the
market is not able to bring about a resource allocation
that is efficient and conducive to solving societal
challenges. A potential reason for that is the existence
of path dependency in technological trajectories as
documented for example in an under-provision of
clean technologies (Aghion et al., 2010). A corollary
of this is that the state has an important coordination
role, helping to remove lock-in effects in
technological developments.
Against this background, this chapter revisits some of
the main arguments in favour of a manufacturing
imperative and discusses them in a European context.
It also shows the limitations and caveats of these
arguments in a world of strong inter-linkages between
the production of manufactures and the services
which enter the production process (Sections: 4.1-
95
Increasing returns to scale can also arise from network
externalities which play a role in a number of sectors that can
be referred to as utilities such as water, gas and electricity,
telecommunication or rail services.
96
Among economists it is highly disputed whether horizontal
measures are necessarily less distortive than sectoral
interventions. De facto, horizontal policies are hardly neutral
with regards to structure and sectors. Therefore, the dichotomy
between horizontal measures and vertical measures may be
blurred or even meaningless (Pelkmans, 2006; Cohen, 2006;
Midelfart and Overman, 2002; Chang, 2006).
4.6). Sections: 4.7-4.10, identify the main challenges
ahead for European manufacturing given the
structural changes that occurred in the EU over the
period 1995 to 2011. Sections: 4.11-4.15 analyse a
number of industrial policy measures that are related
to these structural challenges. Given the still
prevalent use of State aid by EU Member States and
the unique institutional framework which empowers
the Commission to restrict the use of State aid, a
quantitative analysis of State aid and its relationship
with competitiveness and value added is undertaken.
Due to the great importance that the European
Commission attaches to innovation-related industrial
policy, the study of public support measures
continues with a firm-level study of the impact of
public R&D support for firms on innovativeness and
innovation output. The section 4.16 discusses policy
implications of the use of State aid and R&D support
measures in the context of the structural challenges.
4.1. THE MANUFACTURING IMPERATIVE IN A
EUROPEAN CONTEXT
This section lays the ground for the analysis of the
structural shifts in the European manufacturing sector
and the challenges ahead. In particular, it revisits
some of the main arguments in favour of maintaining,
re-building or creating as the case may be a strong
Figure 4.1. Share of manufacturing in value added and in business expenditure on R&D (BERD), 2005-2009
Note: Business Expenditure on R&D includes R&D by foreign enterprises. Averages over the period 2005-2009 of available data.
Source: WIOD, WIPO, OECD ANBERD, wiiw calculations.
manufacturing base in EU Member States while
bearing in mind that modern manufacturing
production is increasingly dependent on innovations
and specialised services inputs. The latter have gained
importance for product differentiation and quality
improvements of manufactures which allow firms to
charge higher prices and increase the value-added of
their activities. Therefore the discussion of the
particular role of manufacturing for the economy has
to be considered in the context of increasing inter-
linkages between manufacturing and services.
Many arguments have been brought forward for why
a thriving manufacturing sector is a prerequisite for
106
any economy aiming for high growth and
employment rates.
4.2. THE MAIN SOURCE OF INNOVATION AND
TECHNOLOGICAL PROGRESS
One principle argument in favour of a strong
manufacturing base is that the manufacturing sector is
the major source of technological progress (e.g.
Baumol, 1967; Kaldor, 1968; UNIDO, 2002;
Aiginger and Sieber, 2006; Helper et al., 2012).
Inspection of firms business expenditure on research
and development (BERD) in the EU and other
countries clearly supports this claim (Figure 4.1).
Manufacturing firms are more inclined to undertake
R&D than firms in the rest of the economy, resulting
in higher shares of the sector compared to its share of
value- added. On average the share of the
manufacturing sector in business R&D exceeds that
of the value-added share by a factor close to four in
the EU Member States; the same holds for the United
States, Japan and South Korea. Despite marked
variations in the business R&D share of
manufacturing firms, ranging from almost 90% in
Germany to 29% in Estonia97, it exceeds the value
added share of manufacturing in all Member States.
Consequently the R&D expenditures of firms indicate
that the overwhelming majority of R&D activities
take place in the manufacturing sector which can
therefore be identified as the main source of
innovation and technological progress.
While the essential role of manufacturing firms for
innovation and technological progress is generally
accepted, an important question is whether a thriving
European manufacturing sector requires innovative
European firms to keep their production facilities in
the EU. For Member States at the technological
97
The median value of the business R&D share of manufacturing
firms is 70.5% for the EU Member States.
frontier it would, in principle, suffice if firms kept
headquarter functions and in particular R&D
activities in the domestic economy but move
manufacturing production to low-wage locations in
order to reduce costs and increase productivity. Such
a vertical specialisation strategy could lead to a high-
powered manufacturing sector in Europe
characterised by highly productive domestically
innovating but internationally producing
manufacturing firms.
While a successful vertical specialisation strategy
supports firms competitiveness and offshoring may
also be seen as a necessity to survive international
competition, a potential risk in this high-powered
manufacturing strategy is a continuous leakage of
more complex activities to offshore destinations. The
stepwise offshoring of more sophisticated production
and engineering activities is the result of the building-
up of capabilities in offshore destinations as well as
communication and co-ordination failures. From a
European perspective, the fact that offshoring is
mainly taking place between EU Member States
could be an advantage, as in this context, and in this
case competences would not risk being shifted out of
the region.
4.3. INCREASED LINKAGES BETWEEN
MANUFACTURING AND SERVICES
R&D and innovation are not the sole ingredients for a
highly productive and internationally competitive
manufacturing sector. In order to differentiate
products and charge higher price-cost mark-ups
manufacturing firms depend increasingly on
sophisticated services inputs. The mirror-image of
this is that the manufacturing sector is an important
source of demand for many services. Both aspects
highlight the fact that goods and services often
complement each other (Nords and Kim, 2013).
Moreover, evidence of the strong interdependences
between manufacturing and services in the European
economy is provided by the fact that manufacturing
firms generate a growing amount of their sales from
services. This servicisation of manufacturing seems
to be more developed among producers of complex
manufactures (Dachs et al. 2013).
On returning to the issue of supply linkages between
services and manufacturing sector, an interesting
indicator is the service intensity of the manufacturing
sector, measured as the cost share of services in
manufacturing gross output. During the period 1995-
2011 the service intensity of the European
manufacturing sector increased from 22% in 1995 to
24% in 2011 with an interim high in 2009 (Figure
4.2).
107
Figure 4.2. Service inputs into the manufacturing
sector relative to manufacturing gross output for the
EU-27, 1995-2011
Figure 4.3. Service inputs into manufacturing
(relative to manufacturing gross output) sourced from
domestic economy, intra-EU and extra-EU, 1995-2011
Note: Calculations based on EU Member States and aggregated
to the EU-27. Intra-EU includes the services sourced from EU
Member States other than the Member States in question.
Source: WIOD, wiiw calculations
This increase, which is discernible in low-tech,
medium-low-tech as well as medium-high-tech
industries, reflects the intensified linkages between
manufacturing and services. It is noticeable that, in
contrast to R&D efforts and innovation which tend to
be concentrated in advanced industries such as
pharmaceuticals, the electronic industry, machinery
and transport equipment industries (particularly the
aircraft industry), there is no systematic relationship
between services intensity and the technology
intensity of industries (see also Nords and Kim,
2013). The reason for this is that transport and sales
services are more intensively used by low-tech
industries. It is true, however, that business services
are most intensively used by the medium-high-
technology industries, although the differences across
the three groups of industries are not very large. This
could mean that precisely because innovation plays a
less important role or international competition is
fiercer, low-tech industries must strongly rely on
business services (such as marketing) in order to
differentiate their products from competitors. An
important feature of the inter-linkages between
manufacturing and services is that EU manufacturing
firms source intermediate services almost exclusively
nationally. On average, the share of domestically
sourced services amounted to 87% in 2011 (Figure
4.3). Another 4% were sourced from other EU
Member States and 9% from third countries.
4.4. THE CARRIER FUNCTION OF MANUFACTURES
Another important structural feature is that
manufactures are highly tradable whereas this is only
true for a subset of services. The higher tradability of
manufactures combined with the increasing services
intensity of manufactures imply that manufactures
assume an important carrier function for services.
Just as many chemical processes require carrier
substances, many services require manufactures to be
carried to foreign customers. This carrier function
stems from the fact that many services by themselves
are not easily tradable as evidenced by the relatively
small (though growing) share of intermediate services
sourced from abroad. The high tradability of
manufactures and the carrier function this provides
for services are of course highly relevant for the EUs
external balance of payments.
While the share of services in the EUs gross exports
to third countries has grown considerably over the
past decades to about a third, it still falls far short of
the (equally growing) share of services in both GDP98
and value-added exports99. This can be seen by
comparing the share of services in gross exports, i.e.
33%, to the share of services in extra-EU value added
exports which amounted to 57%. Hence, in terms of
value-added exports the share of services exceeded
that of manufactures which amounted to 37% in
2011. The rising importance of services in terms of
value-added exports results from the fact that more
services are embodied in exports of the
manufacturing sector than vice-versa100. Hence, for
non-tradable services an internationally competitive
manufacturing sector is needed in order to make
services exportable and to create comparative
advantages in services101. At the same time, services
have become an essential factor in underpinning the
competitiveness of manufactures.
4.5. PRODUCTIVITY GROWTH
Another common argument for the special role of
manufacturing which is strongly related to the
innovation argument but distinct from it is that
productivity growth is higher in manufacturing than
in the rest of the economy. The productivity argument
98
Typically, the share of services account for about 60-70% of
GDP in advanced economies.
99
Value added exports are a measure based on input-output
methodology that reflects the value-added created domestically
in an industry or sector in order to satisfy foreign demand (see
also Box 4.1).
100
Another factor is that vertical specialisation and trade in
intermediates in general is more developed in manufacturing
which inflates the gross amounts of exports.
101
An alternative way to sell services internationally is by
establishing a foreign subsidiary (Mode 3 of cross-border
services trade in WTO terminology).
108
is related to the innovation argument because R&D
and innovation feed into technological progress and
productivity growth. It is distinct because the sector
of origin of technological progress need not
necessarily be the sector that benefits most strongly
from new technologies102.
Irrespective of this distinction, it turns out that total
factor productivity (TFP) growth in the
manufacturing sector outperforms TFP growth in the
total economy as well as that of business services
across a sample of EU Member States and also the
US (Figure 4.4). Within the EU, the TFP growth
differential between the manufacturing sector and the
total economy is particularly large in Austria and in
Germany, but is also present in the service-oriented
British economy. The sole exceptions to this EU-wide
pattern are Spain and Italy which actually did not
experience any TFP growth between 1995 and 2007.
The result remains unchanged if TFP growth in
manufacturing is compared to TFP growth in the
market services sector instead of the total economy.
Hence, the superior TFP growth trajectory in the
manufacturing sector between 1995 and 2007 is not
due to low productivity performance in typically low
productivity services such as health care or personal
services.
TFP growth in the manufacturing sector also exceeds
that of the total economy in the United States103.
The reason for higher productivity growth in the
manufacturing sector is partly related to technological
aspects of manufacturing (increasing returns to scale,
externalities, learning effects)104. An additional
reason is that manufactures, being more tradable
compared than services, are exposed to fiercer
international competition which sets further
incentives to increase productivity. This does not
exclude the possibility of high productivity pockets
within the services sector which is of course a very
heterogeneous sector, comprising a number of high
productivity industries like telecommunications.
102
The relationship between innovation and productivity at the
industry or sectoral level is blurred by the fact that in the case
of product innovations the productivity gains (depending on
market structures) may not accrue to the innovating industry
but to downstream industries sourcing cheaper inputs or inputs
of higher quality. By contrast, productivity gains from process
innovation typically accrue in the innovating sector though
they may spread to other sectors later on.
103
In the case of the United States, however, real productivity
growth of manufacturing may be overstated due to strongly
decreasing price deflators in the electronic equipment industry.
104
Another issue is the problem of measuring and comparing TFP
across industries, but lacking alternatives this analysis relies on
the best data source available which is the EU KLEMS
database.
Figure 4.4. Comparison of total factor productivity
(TFP) growth in the manufacturing sector, the total
economy and market services, 1995-2007
Source: EU KLEMS, wiiw calculations.
An implication of these differentiated patterns of TFP
developments is in accordance with Baumols
arguments of structural change (see Baumol, 1967)
outlined in more detail below - that in the longer term
prices of manufactures will decline relative to
services which ceteris paribus - is leading to a lower
share of manufactures in value added in nominal
terms. Therefore a declining value added share of the
manufacturing sector per se is not a reason for
concern but the logical consequence of a European
manufacturing sector that is constantly becoming
more efficient.
To sum up, the comparison of TFP growth rates
supports the view that the manufacturing sector is not
only the most important source of innovation and
technological progress but also the sector where
innovations and new technologies are primarily
implemented and turned into total factor productivity
growth.
4.6. DOES MANUFACTURING OFFER HIGHER WAGES
IN EUROPE?
A final argument in the context of a manufacturing
imperative is that the manufacturing sector is capable
of providing a large amount of well-paid jobs
(Rodrik, 2012). This claim is typically put forward in
the context of emerging economies but it could also
be relevant for the cohesion countries among the EU
Member States.
From a theoretical perspective, the argument that the
manufacturing sector offers higher wages typically
states that the production of manufactures is
characterised by imperfect competition (e.g. due to
learning effects or static economies of scale in
production), combined with imperfect inter-industry
labour mobility within a country105. For the EU-27,
105
From a theoretical perspective differences in wages between
industries will always depend on some limitations to inter-
109
however, there is no evidence of higher wages in
manufacturing compared to the services sector
neither at the general wage level, nor for wages set by
educational attainment. Considering the EU as a
whole, hourly wages have been lower in the
manufacturing sector (EUR 13.39) than in the
services sector (EUR 14.34)106. At the level of EU
Member States the results are mixed, with
manufacturing wages being higher in some EU-15
countries. But wages in the services sector are higher
in all central and eastern European Member States as
well as Malta and Cyprus (EU-12). The same
comparison but taking the educational attainments of
workers into account, suggests that in general wage
differentials between the services and the
manufacturing sector are small. The finding is in line
with the results found for other countries, such as the
United States (McKinsey Global Institute, 2012).
According to economic theory, factor rewards should
in the long run reflect factor intensities.107 Simple
correlations between wages in different sectors could
therefore be misleading.
4.7. STRUCTURAL CHANGE IN THE EU ECONOMY
A general feature of the European economy (and
advanced economies in general) is the structural shift
to the services sector. This shift is observable for both
value-added and employment and has been discussed
in Chapter 2 of this report. The mirror image of the
move into services in Europe is a decline in the
relative importance of manufacturing industries
(Table 4.1) for which there is a whole series of
explanations.
As shown above, productivity growth in the European
manufacturing sector outpaces productivity growth in
services and the economy in general.
This is a major reason why relative prices of
manufactures decline relative to those of services. As
a consequence, the nominal value added share of
manufacturing declined by 4.2 percentage points
between 1995 and 2011 (and by 5.3 percentage points
between 1995 and 2009) as shown in Table 4.1. The
relative decline in real terms was more moderate,
amounting to 2.6 percentage points between 1995 and
2009 (see for example also Aiginger, 2007). In real
terms, the value added share of the EU manufacturing
sector is higher than in nominal terms amounting to
17.5% in 2009. The share of the manufacturing sector
in terms of employment declined to a similar extent
industry labour mobility. Differences in wages can be
motivated by a number of economic models, e.g. a specific-
factor model of trade. The differences in wages between
industries depend on a number of factors including the capital
intensity or whether one looks at the short or the long run.
106
This result is based on 2010 Eurostat data of hourly gross
earnings of employees working in companies with ten or more
employees.
107
Norman, V, D. & Orvedal, L. (2010).
as the nominal value added share (4.3 percentage
points between 1995 and 2009).
This suggests that technological progress which lies
behind the changes in relative prices is mainly labour-
saving.
The structural shift out of manufacturing (both in the
EU and globally) encompasses basically all
manufacturing industries, implying that the aggregate
decline of the manufacturing value-added share is the
result of widespread trends across industries rather

Table 4.1. Nominal and real valued added shares and employment shares in the EU and the global economy 2009 and
2011 (in %); changes 1995-2009 and 1995-2011 in percentage points
EU-27 World
Nominal value Real value added Employment Nominal value Real value added Employment
added added
Industry 2011 change 2009 change 2009 change 2011 change 2009 change 2009 change
1995- 1995- 1995- 1995- 1995- 1995-
2011 2009 2009 2011 2009 2009

Primary Industries 2.7 -1.21 3.1 -0.79 5.9 -3.73 9.6 3.29 4.9 0.22 32.2 -8.76
Manufacturing 15.8 -4.24 17.5 -2.55 15.6 -4.33 17.2 -2.43 18.3 -1.53 15.2 0.20
Food 1.9 -0.54 2.0 -0.45 2.2 -0.46 2.4 -0.20 2.1 -0.40 1.9 -0.20
Textiles 0.5 -0.55 0.6 -0.43 1.1 -1.04 0.8 -0.27 0.8 -0.25 2.6 0.29
Leather 0.1 -0.09 0.1 -0.11 0.2 -0.20 0.1 -0.02 0.1 -0.04 0.5 0.15
Wood 0.3 -0.15 0.4 -0.11 0.6 -0.21 0.4 -0.13 0.3 -0.15 1.0 0.27
Pulp & Paper 1.2 -0.64 1.5 -0.38 1.1 -0.42 1.1 -0.53 1.3 -0.37 1.0 0.22
Ref. Petroleum 0.3 0.00 0.3 -0.04 0.1 -0.06 0.9 0.27 0.7 0.04 0.1 -0.02
Chemicals 1.7 -0.39 2.2 0.12 0.8 -0.30 1.8 -0.17 2.0 0.06 0.8 -0.11
Plastics 0.7 -0.20 0.9 0.00 0.8 -0.04 0.7 -0.15 0.7 -0.11 0.9 0.28
NM Minerals 0.6 -0.34 0.7 -0.24 0.7 -0.23 0.7 -0.15 0.7 -0.19 0.9 -0.37
Metals 2.4 -0.29 2.2 -0.53 2.3 -0.40 2.4 -0.23 2.2 -0.48 1.3 -0.24
Machinery 2.0 -0.14 1.9 -0.30 1.7 -0.42 1.5 -0.20 1.7 -0.14 1.1 -0.19
Electrical Eq. 1.7 -0.56 2.6 0.27 1.7 -0.30 2.3 -0.18 3.3 0.78 1.4 0.22
Transport Eq. 1.7 -0.18 1.8 -0.16 1.4 -0.13 1.6 -0.36 1.9 -0.16 0.9 -0.01
Manufacturing n.e.s. 0.6 -0.17 0.6 -0.20 1.0 -0.12 0.5 -0.11 0.5 -0.12 1.0 -0.09
Electricity, gas, water 2.4 -0.29 2.2 -0.48 0.8 -0.25 2.1 -0.20 2.2 -0.28 0.5 -0.02
Construction 5.9 -0.10 4.8 -1.19 7.2 0.16 5.5 -0.38 4.4 -1.48 6.9 1.36
Services 73.2 5.84 72.4 5.01 70.5 8.15 65.6 -0.29 70.1 3.07 45.1 7.22
Note: Industry classification based on NACE Rev. 1.1. Food=15t16; Textiles=17t18; Leather=19; Wood=20; Pulp & Paper=21t22;
Refined Petroleum=23; Chemicals=24; Plastics=25; Non-Mineral Metals=26; Metals=27t28; machinery=29; Electrical
equipment=30t33; Transport equipment=34; Manufactures n.e.s.=36t37. World includes EU-27.
Source: WIOD, wiiw calculations.

A second factor for the observable structural trend is
rigid demand structures characterised by low price
elasticities of demand and high income elasticities for
some services, e.g. education, tourism, health,
cultural activities (see Baumol, 1967). This factor
helps to explain why the relative importance of
manufacturing in value added terms has declined over
time. Besides, outsourcing processes and vertical
disintegration (with more service activities provided
by external firms rather than produced internally)
might be an additional cause for declining industry
shares.
than driven by only a part of them.
Against the background of these general structural
trends at the global and European level, important
changes in the global economy such as the emergence
of new players in international production and trade
and the growing importance of ideas, skills and
technology for international competitiveness, poses
major challenges for European manufacturing.
To offset the effects of cyclical consumption patterns
and sectoral relative productivity growth rates on
manufacturing shares of GDP and total employment,
manufacturing firms and industries in the EU need to

110
become more competitive on the world markets.
Given the importance of service inputs in
manufacturing production, the completion of the
single market for services is expected to advance the
level of services tradability. By raising their market
shares, the production and employment in EU
manufacturing can increase.108
Industrial policy also has a role to play here, by
providing the rules and instruments necessary to
increase the competitiveness of EU manufacturing
industries.
mentioned before tends to be labour-saving110. In
addition the structural shifts within the manufacturing
sector are going in the direction of a mild but
persistent shift towards more technology-intensive
industries (chemicals, machinery, electrical
equipment and transport equipment) which also tend
to be less labour-intensive. These advanced
industries also registered negative employment
trends between 1995 and 2009 (with the exception of
transport equipment) but job losses were more
pronounced in the low-tech industries (3.5 million)
which accounted for 70% of total losses in

Table 4.2. Employment developments within the manufacturing sector, EU-27, 1995-2009
1995 2009 changes 1995-2009

number of jobs number of jobs number of jobs percentage

industry share share
(in '000) (in '000) (in '000) points

low-tech 17,257 43.1 13,795 39.3 -3,462 -3.78

medium-low tech 3,778 9.4 3,493 10.0 -285 0.52

metals 5,419 13.5 5,155 14.7 -264 1.16

chemicals 2,258 5.6 1,864 5.3 -394 -0.33

machinery 4,227 10.6 3,786 10.8 -441 0.23

electrical eq. 3,958 9.9 3,758 10.7 -200 0.83

transportation eq. 3,142 7.8 3,235 9.2 93 1.37

manufacturing 40,038 100.0 35,084 100.0 -4,954

Note: Value added price deflators for the electrical equipment industry of Finland, France, Sweden, Japan, South Korea and the US
replaced by respective German deflation in each year. Industry classification based on NACE Rev. 1.1. Low-tech: Food=15t16,
Textiles=17t18, Leather=19, Wood=20, Pulp & Paper=21t22, Manufactures n.e.s.=36t37; medium-low-tech: Refined Petroleum=23,
Plastics=25, Non-metallic mineral products=26; Metals=27t28; Chemicals=24; Machinery=29; Electrical equipment=30t33; Transport
equipment=34;
Source: WIOD, wiiw calculations.

manufacturing employment (Table 4.2. ).

4.8. TRENDS WITHIN EU MANUFACTURING
Over the period 1995-2009 almost 5 million jobs
were lost. From 2009 to 2011 manufacturing
employment in the EU-27 fell by another 1 million
jobs109.
To some extent the loss of manufacturing jobs may be
offset by new jobs created in services sectors
providing intermediate services to manufacturing.
An important explanation for the negative
employment developments in European
manufacturing is the increase in productivity that as
This trend towards advanced manufacturing
industries reflects international specialisation patterns
of EU Member States because in general technology-
intensive industries offer more possibilities for
building comparative advantages by product
differentiation and quality aspects. At the same time
low-technology-intensive industries still accounted
for almost 40% of manufacturing employment in
2009. Overall, the EU manufacturing sector is well
diversified. In order to maintain diversity, the
structural upgrading should proceed at a moderate
pace in order to ensure that the manufacturing base in
the EU remains broad, encompassing all industries. In
low-tech and medium-low-tech industries this will

108 110
The effects on employment could however be negative if the This has to be considered in conjunction with the structures of
main way to become more competitive is to increase price and income elasticities of demand which tend to work
productivity growth. against compensating demand shifts towards relatively cheaper
109
Development 2009-2011 based on Eurostat data. manufactures.

111
require a high degree of specialisation within these
industries and the occupation of niche markets.
Existing evidence suggests that many European firms
follow such a premium strategy within their
respective industry. Within industries and product
categories featuring a low degree of complexity,
European firms typically operate in the top quality
segments (Reinstaller et al., 2012)111.
Maintaining a broad and well-diversified
manufacturing base in Europe is important in order to
preserve manufacturing capabilities which, once lost
are hard to develop again. Manufacturing capabilities
specific to particular industries may at a later stage
turn out to be important inputs for fast growing new
products. It is argued that the United States has made
this experience in several industries such as shoe
production where the entire supply chain has been
lost (Helper et al., 2012).112
Having stressed the diversification of the EU
manufacturing sector and the specialisation into the
premium segments within industries it is also
important to note the high degree of heterogeneity
across Member States. Figure 4.5 illustrates this with
respect to the value added share of manufacturing and
changes thereto between 1995 and 2011. While this is
an imperfect indicator of the role of the
manufacturing sector for the economy, the cross-
country comparison still indicates which countries
may have reason to be worried about their industrial
commons. There is cause for concern either because
the value added share of manufacturing is declining
very strongly as in the United Kingdom or Latvia
or because it was very low initially (i.e. in 1995) as in
the case of France or Greece.
In principle, a declining share of manufacturing in the
economys value-added may be of little concern in
Member States whose manufacturing industries
produce sophisticated products with a high premium
on world markets. Such Member States, e.g. Finland
(despite some problems faced recently in the high-
tech manufacturing sector) or the United Kingdom
are potentially left with a high-powered
manufacturing sector. The developments could be of
more concern in countries such as Cyprus, Greece,
Latvia or Malta where industries produce less
sophisticated products and are therefore more
vulnerable to competition from low-cost producers.
(see Reinstaller et al., 2012). In contrast, there is a set
of countries including Germany, Austria and a
number of central and eastern European countries that
have maintained a rather high-value added share of
manufacturing. This shows that there is considerable
dispersion among Member States.
4.9. EXTERNAL COMPETITIVENESS
Figure 4.6. not only shows the competitive positions
of the EU-27 as measured by shares in global value-
added exports compared to major competitors among
advanced economies the US, Japan and South
Korea but also compared to large emerging
economies like Brazil, China and India (for the
concept of value-added exports see Box 4.1).
Technological leadership and quality upgrading have
become increasingly important to ward off
competition from emerging economies.
Given the structural upgrading in emerging
economies, competitive pressures from these
countries are not limited to low-technology-intensive
industries but are also felt in advanced manufacturing
industries where emerging economies have also
gained a foothold. Brazil, India and China all
considerably increased their market shares in global
value-added exports of manufactures. However, it is
the outstanding performance of China, whose market
share quadrupled between 1995 and 2011, which
basically drives the reshuffling of competitive
positions in the global economy.
By 2011 China had almost caught up with the EU-27
in terms of value added exports of manufactures, with
both economies having a market share of about 20%.
Chinas rise to a first-class exporter of manufactures
can be seen in the way it gained export market shares
across all industries, with extremely strong positions
in the export of textiles and leather as well as in the
electrical equipment industry. While China is still
specialised in the relatively more labour-intensive
stages of production within the electrical equipment
industry, the impressive gains in market shares also
reflect a remarkable upgrading of industrial
structures. The same holds true for other industries
and also other emerging economies, e.g. the Indian
pharmaceutical and automotive industries.
A factor that facilitated structural upgrading in
emerging economies is the relative ease of
international technology transfer in a global economy
(through trade, FDI, labour mobility in the high-skill
segment of the labour force and knowledge
diffusion). This is particularly true for the
manufacturing sector because the required technology
and industrial know-how are to a large extent
embodied in physical products which makes them
more prone to imitation.

111
See also Chapter 2 of this Report.
112
Maybe more important is the case of thin-film-deposition
which has moved from the US to South East Asia together
with semiconductor production which turned out to be
important for producing solar panels.

112
Figure 4.5. Developments of the value-added share of manufacturing (nominal) across EU Member States and selected
competitor countries, 1995-2011
35
CN

IE
30

KR
SK
share of manufacturing 1995

SI
RO
25 FI
CZ
SE JP DE
MT IT
ES HU
LV BG PL
UK BE
20 AT
IN LT
ES
PRT BR

DK NL

US
15
FR
LU

CY EL

10
-15 -10 -5 0 5
change in value added share of manufacturing 1995-2011

Source: WIOD, wiiw calculations.

Figure 4.6. Shares in global value added exports of manufactures (in %), 2011 (upper panel) and changes thereto (in p.p.),
1995-2011 (lower panel), extra-EU exports

EU-27 US JP KR BR CN IN shares in global

41.3 46.0 value added exports, 2011
30
market share in %

20

10

0
Ref. Petroleum
Leather

Wood

Plastics
Textiles

Metals
Chemicals
Food

Machinery
NM Minerals
Pulp & Paper

Electrical Eq.

Manufacturing

Manufacturing
Transport Eq.

nec.

23.3 27.6
20
change in shares in p.p.

10

-10 changes in shares, 1995-2011

-20

Note: Industry classification based on NACE Rev. 1.1. Food=15t16; Textiles=17t18; Leather=19; Wood=20; Pulp & Paper=21t22; Refined
Petroleum=23; Chemicals=24; Plastics=25; Non-metallic mineral products=26; Metals=27t28; Machinery=29; Electrical equipment=30t33;
Transport equipment=34; Manufactures n.e.s.=36t37. Global market shares in value-added exports and changes thereto exclude intra-EU value-added
exports. Source: WIOD, wiiw calculations

113
Box 4.1. Why is it important to look at value-added exports?

International trade has not only expanded spectacularly over the past 25 years, it has also become increasingly
complex. One important dimension in this complexity is the fact that the specialisation patterns have become
more granular. Supported by declining trade costs, the ever finer specialisation on individual components of a
product or steps in the production process also referred to as fragmentation of production makes the
analysis of trade flows more demanding. International fragmentation of production heightens the importance
of trade in intermediate goods. This in turn poses some difficulties for traditional trade statistics which record
trade flows according to a gross concept thereby inflating trade figures.

One possibility to adjust gross export flows for imported intermediates is provided by global input-output
statistics. The present Report relies on the World Input-Output Database (WIOD) which provides such
statistics for a set of 40 countries including EU Member States. The WIOD is used to calculate the value-
added exports at industry level for each country or country groups. These value added exports capture only
the value added that is generated domestically in the production of goods that are destined for export (see
Johnson and Noguera, 2012; Stehrer, 2012) but exclude foreign value-added associated with imported
intermediates.

The figures above illustrate how the differences between gross exports and value-added exports can be quite
significant, particularly in industries characterised by intensive intra-industry trade such as electrical
equipment. According to gross exports, Chinas market share in the electrical equipment industry for example
rose from 5.27% in 1995 to 33.6% in 2011. Looking at value added exports, Chinas market share still rose
but reached only 24.5% in 2011. While this is still a spectacular development, the resulting difference
between Chinas market share in gross exports and value added exports is equal to about 33.6% and 24.5% in
2011, respectively.

For the EU and the United States the opposite is true. The EUs share in global value added exports in the
electrical equipment industry is 2.2 percentage points higher than in terms of gross exports in 2011 and in the
US the difference even reaches 9 percentage points. The figures above also indicate that the difference
between gross exports and value added exports has increased between 1995 and 2011 due to the emergence of
international production networks and more fragmented global production.

In the presence of international production sharing the value added exports probably give a more accurate
picture of export market shares of the trading partners involved

As a result emerging economies like China not only
have large export market shares in low-tech and
medium-low-tech industries (where they can be
expected to possess comparative advantages due to
lower labour costs) but also increasingly in more
technology-intensive industries.
The mirror image of the entry of China and other
emerging economies into the global trade arena is a
decline in market shares of the EU, the US both lost
about a fifth of their export market shares between
1995 and 2011 and Japan, whose market share was
halved. Even if the gains in market shares of China
levels off in coming years as wages rise and the
technology gap narrows113, these industries in China
will remain major competitors. Arguably, competition
may become fiercer as the catch-up process of major
emerging economies such as China, India and Brazil
continues and these countries expand their skills and
capabilities in the manufacturing domain.
In any case, the shifts in competitive positions
discernible in Figure 4.8. suggest that the EUs losses
of export market shares in manufacturing were
primarily due to the integration of emerging countries
into the global economy and to a lesser extent to
competition from other advanced economies with the
exception of South Korea, which made substantial
inroads into the production and export of transport
equipment (mainly the automotive industry).
Given these trends in market shares in global value
added exports, further shifts towards these emerging
economies can be expected.

113
Gains in market shares in Chinese value added exports in
manufactures seems to have levelled off somewhat since the
mid-2000s although they continued to increase (by 4.2
percentage points between 2007 and 2011 compared to 5.3
percentage points between 2002 and 2006).
From a European perspective, however, the rise of
China and other emerging economies not only
constitutes a formidable competitive challenge but
also means new and enlarged markets. Equally
important is the fact that the benefits from new export

114
opportunities and the potential costs of a deteriorating
international competitiveness are not distributed
evenly across EU Member States. This leads to
another main challenge for European manufacturing
which consists of the agglomeration of manufacturing
activities.
Figure 4.7. Differences between market shares in gross exports and value added exports in the electrical equipment
industry, 1995-2011
Source: WIOD, wiiw calculations.
4.10. R&D AS MEANS TO MEET COMPETITION
The gap between the innovation activities of firms in
the EU and the United States has been a concern for
European policymakers for decades. Indeed, the
comparison of R&D intensity in manufacturing as an
indicator of the intensity of innovative activity,
measured as the business expenditure of
manufacturing firms on R&D relative to
manufacturing value-added, suggests that European
manufacturing is characterised by lower R&D
intensity in comparison to US and Japan.
Figure 4.8. Decomposition of differences in manu-
facturing R&D intensity in EU Member States, the US
and Japan, average 2007-2008
composition effect
interaction effect
2,0
deviation from global average
0,0
-2,0
-4,0
-6,0
AT BE DE FI
Note: R&D intensity is Business expenditure on Research and
Development in per cent of value added. Global average is the
average of the nine countries. R&D intensity differential is the
difference of the manufacturing-level R&D intensity to the mean
of the nine countries. Methodology following Eaton et al. (1998).
Industry classification based on NACE Rev. 1.1. For industry
groupings and decomposition see Appendix
Source: WIOD, OECD ANBERD, wiiw calculations
These differences in R&D intensity at the
manufacturing level can be split into a composition
effect which reflects differences across countries in
industry structure, and an intensity effect which
reflects differences in the R&D intensity at the level
of manufacturing industries, as well as an interaction
effect (see Eaton et al., 1998 and European
Commission (2011)). This decomposition shows that
the differences in the R&D intensity of firms across
EU Member States, the US and Japan at the
manufacturing level are mainly driven by the
intensity effect. The industry structure (composition
effect) plays a role in some Member States but is
never the primary factor114.
This gap in R&D activities of the manufacturing
sector in the seven EU Member States, characterised
by the largest R&D intensities across Member States
for which data are available, is partly compensated by
higher public R&D expenditure in these countries.
R&D intensity in the seven EU Member States with
the relatively highest R&D intensities across the EU
intensity effect
(Figure 4.8) is only 62% that of the United States.
R&D intensity differential
However, there are also research findings-based on
BERD territorial official statistics and also company
data that conclude on the lower overall corporate
R&D intensity for the EU as a result of sector
specialisation (structural effect). In these cases, the
US seems to have a stronger sectoral specialisation in
the high R&D intensity (especially ICT-related)
114
FR UK NL US JP
The relative importance of the composition effect and the
intensity effect in such a decomposition exercise depends on
the level of aggregation of the industries. A more detailed
industry break-down would assign greater importance to the
composition effect. The EU R&D Scoreboard 2012 also
identifies an R&D intensity gap which is particularly strong in
the high-tech industries when using ANBERD data which are
territory based. When using a more elaborate analysis based on
company-level data, it is argued that industry composition
becomes an important agent (see
http://iri.jrc.ec.europa.eu/scoreboard12.html for details,
particularly Chapter 7).
115
sectors than the EU does, and also has a much larger
population of R&D investing firms within these
sectors115. This issue, call for policy makers to pay
further attention to the industrial structures
differences and the need for Europe to favour the
growth and emergence of new world leading
innovative companies116.
At the same time, it seems that the concern about a
deterioration of relative positions in advanced
manufacturing industries vis--vis the US and other
economies at the technological frontier; should be
limited to the electrical components industry. In all,
other advanced manufacturing sectors, the market
shares in global value-added exports of the EU are
still much higher than those of the US. The EU is still
the worlds largest exporter of chemicals, machinery
and transport equipment, with the latter two
constituting the major strongholds of European
manufacturing. Despite a 6 percentage points decline
in its market share of global value added exports
between 1995 and 2011117, the EU still accounts for
more than a third of global machinery valued-added
that is exported, putting it far ahead of the United
States118. The EU-27 also has considerable export
market shares in low-technology industries such as
the food industry or the pulp and paper industry
which supports the claim that EU firms often occupy
premium segments within industries to remain
internationally competitive. An example for such
high-quality specialisation in low-technology sectors
is the production of protective textiles or extra-long
hardened rail tracks. Figure 4.8 suggests that EU
firms are more successful in this type of
specialisation than their US rivals.
Offshoring implies that part of the value-added
created by EU firms is generated in low-cost
locations. The offshoring activities of EU
multinationals were predominantly regional in scope,
meaning that labour-intensive parts of the production
process were re-located to central and eastern
European Member States, which also still have
relatively low labour costs by EU standards. It is
worth mentioning that offshoring does not
predominantly affect labour-intensive industries (as
opposed to advanced manufacturing industries). The
dividing line is rather the skill level of employees
with low-skill (though often medium-paid) jobs in
manufacturing being more prone to offshoring. This
points towards a major role for education and training
115
See Moncada-Patern-Castello; Ciupagea, C.; P., Smith, K;
Tbke, A. and Tubbs, M.: "Does Europe perform too little
corporate R&D? A comparison of EU and non-EU corporate
R&D performance", Research Policy 39 (2010) pp. 523536
116
See Cincera, M. and Veugelers, R.: Young Leading Innovators
and EUs R&D intensity gap. JRC-IPTS Working Papers on
Corporate R&D and Innovation, n11/2010.
117
These figures exclude intra-EU value added exports.
118
These figures exclude intra-EU value added exports.
116
of the labour force, in particular in high-wage
countries, in order to remain an attractive location for
manufacturing activity.
The next section will investigate the effects of
industrial policies in the form of State aid and
innovation support.
4.11. INDUSTRIAL POLICY MEASURES IN THE
EUROPEAN UNION
Few people will doubt that the main responsibility for
mastering the challenges facing the European
manufacturing sector rests with firms. However, a
recent survey of the top 1000 EU R&D investing
companies has shown that public policies may
constitute an important stimulus for company
innovation119. According to that survey, national
public support in terms of fiscal incentives and public
grants had a positive effect on company innovation,
as well as EU policies in terms of direct public aid
and public private partnerships. In this sense, another
question is whether the EU and its Member States
have fully exploited the potential of industrial policies
to support firms in mastering these challenges and
ensuring a strong manufacturing base in Europe.
After a brief overview of industrial policies at the
Union level and by Member States, this section
provides a quantitative analysis of State aid by
EU Member States which constitutes an important
industrial policy tool. Since R&D is a key aspect in
EU's industrial policy mix and it is directly linked to
the challenges of European manufacturing, this
section also investigates the impact of public funding
for R&D on innovation activity and innovation output
at firm level.
4.12. INDUSTRIAL POLICIES AT THE UNION LEVEL
AND BY MEMBER STATES
With the Maastricht Treaty, the EU enshrined its
industrial policy approach in primary law, stipulating
that the Union and the Member States shall ensure
that the conditions necessary for the competitiveness
of the Unions industry exist120. However, by
defining industrial policy in a very broad sense
including flanking measures, the EU had set a major
industrial policy objective well before Maastricht.
The creation of the European single market embedded
the EUs competition rules, which were previously
part of the earlier common market. The particularity
of EU competition rules is that besides controlling the
anti-competitive behaviour of firms (abuse of a
dominant position, market and price rigging and later
119
See Tbke, A.; Hervs, F. and Zimmermann, J.: "The 2012 EU
Survey on R&D Investment Business Trends", European
Commission, Joint Research Centre, EUR 25424 EN,
www.jrc.es, pp.22.
120
Article 173(1) of the Treaty on the Functioning of the
European Union (TFEU).
merger control), the European Commission was also
empowered to control the State aid provided to firms
by EU governments. Still today this is a quite unique
feature in competition rules.121 The control of State
aid of sovereign governments is obviously a delicate
issue and the Commission has exercised a large
degree of pragmatism in this respect (Doleys, 2012).
The Commission has tried to shift State aid by
Member States from sector-based schemes to
horizontal objectives such as aid to SMEs or support
for R&D, or aid for employment and worker training.
The European Commissions preference for
horizontal State aid stems from the belief that
horizontal aid distorts competition less than sectoral
aid (Friederiszick et al. 2006), and that it contributes
to the Commissions own market-correcting or
redistributive policy goals which links it to an
objective of common interest (Blauberger, 2008)122.
The latest revision of State aid policy (State Aid
Modernization) favours a shift towards block-
exempted aid. This aid is less likely to distort
competition, due to lower levels of aid intensities and
it is in line with the most prominent EU initiatives.
While at the Union level, the focus remained on
general framework conditions, there were also early
attempts to implement a kind of technology policy
(Owen, 2012). Over time, the support for R&D,
innovation and technology funded from the EU
budget has become quite substantial, leading
prominent economists to conclude that at the EU
level industrial policy is essentially R&D policy (Van
Pottelsberghe, 2007).
The EUs ambitions in the field of industrial policy
(European Commission 2010a) have intensified in the
aftermath of the economic crisis of 2008 with the
focus largely remaining on framework measures and
innovation. Hence, in the EUs new growth strategy,
the Europe 2020 strategy adopted in 2010 (European
Commission, 2010b), the Innovation Union
(European Commission, 2011c) figures prominently
among the flagship initiatives. Moreover, the 2020
strategy also confirms the EUs horizontal industrial
policy approach. This policy communication also
proposes a fresh approach to industrial policy that
121
Only EFTA has a comparable competition authority.
122
For the various types of horizontal state aid there exist so-
called block exemptions. These block exemptions specify a
number of criteria that aid programmes must fulfil (e.g.
maximum subsidy amount typically expressed in percentage of
eligible costs). If the criteria are fulfilled the aid programme is
considered to be compatible with state aid rules. The block
exemptions constitute a major simplification of the state aid
procedure as they exempt eligible aid programmes from the
requirement of prior notification and Commission approval.
For Member States this means that they are able to grant aid
that meets the conditions laid down in these regulations
without the formal notification procedure. However, ex post
information sheets on the implemented aid have to be
submitted.
117
complements its market-oriented horizontal approach
with sector-specific elements. The Commission
characterises its approach as bringing together a
horizontal basis and sectoral applications (European
Commission, 2010a, p. 4). The mention of sectoral
application of horizontal measures seems to take into
account the claim that infrastructure and other public
inputs tend to be highly context-specific, calling for a
sector-specific definition of industrial policy
(Hausmann and Rodrik, 2006). In the specification of
the sectoral dimension of industrial policy, the
European Commission identifies the development of
clean and energy-efficient vehicle technologies as a
priority area for industrial policy. The Commissions
update of the industrial policy communication from
October 2012 (European Commission, 2012a)
contains six priority action lines which aim at
improving the competitiveness of European
manufacturing.
These priority lines highlight once more the
importance of new technologies for a thriving
manufacturing sector. At the same time these action
lines are directly or indirectly related to the protection
of the environment and the mitigation of climate
change.
The priority action lines are accompanied by a
number of additional objectives, such as the
establishment of a European patent, and new elements
such as the green public procurement, a demand-
side policy instrument which has not previously
featured among the main concerns of industrial
policy.
The industrial policy approach at Union level is
highly relevant for the industrial policies applied by
Member States. The interdependence between
policies at EU level and Member State level is most
obvious in the field of competition policy including
State aid where the Commission is responsible for
controlling the activities of Member States. But the
two layers are also linked by the fact that most of the
projects paid from EU funds have to be co-financed
by Member States.
During the 1980s, State aid to industry and services
provided by EU Member States amounted to
approximately 2% of EU GDP and went down to
about 1% in the following decade (European
Commission, 2011b). The general downward trend in
State aid in the EU continued until 2007 where it
reached an all-time low of 0.4% of GDP.
State aid also has a counter-cyclical component, i.e.
the amount of state aid spent increases in times of
recessions. This was the case in the economic crisis
of 2008. As shown in Figure 4.9., state aid increased
to 0.6% in 2008, which is still a very low amount by
historical standards but represented a 50% increase
from the year before. These figures include state aid
granted under the Temporary Framework which state aid provided by Member States has become
allowed a temporary adjustment of state aid rules and relatively small. Secondly, the renewed interest in
was intended to encourage investment and ease the industrial policy both at the Member State and the EU
access to finance for firms facing tightening credit level has not so far resulted in a substantial increase
conditions. It was targeted at the real economy.123 in state aid figures126. Thirdly, the impact of even
small amounts of state aid is potentially very large.
The Temporary Framework provided new measures The total of state aid measures under the Temporary
specifically targeted to facilitate companies access to Framework by the 27 Member States sums up to EUR
finance124. 4.8 billion over the period 2008-2010 but it consists
of a large number of measures, including multi-
Figure 4.9. State aid to industry and services in the EU- billion loans to car producers. The aid elements
27, 1992-2011, in % of GDP implicit in such measures seem low but they can
Temporary Framework
nevertheless have a great impact on individual
non-crisis state aid companies (in particular when the state aid comes in
GDP growth in % (right axis) the form of a rescue operation) but also at market
EU GDP 4 level for the industry127. So the leverage of state aid
growth
1,0
measures may be quite high. EU governments have a
2
GDP growth in% great potential to affect market outcomes and also the
position of EU companies in global competition
state aid in% of GDP

0
without large fiscal implications.
0,5 -2 The next sections analyses use of state aid by EU
Member States in more detail by investigating the
-4 relationship between various types of state aid on the
one hand and competitiveness and value added of the
0,0 -6 manufacturing sector on the other hand.
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011

Note: Figures exclude crisis-related aid to the financial sector.

The value for France in 1997 excludes the EUR 18 billion state
aid to Crdit Lyonnais. Amounts refer to the aid element (or
gross grant equivalent in the case of guarantees and loans)
contained in the state aid measure.
Source: European Commission State Aid Scoreboard, Eurostat,
wiiw calculations.

Due to the crisis, state aid by Member States rose to

0.5%-0.6% of GDP in the years 2008-2010 but in
2011 the amount dropped to 0.44% of EU GDP,
which equals the pre-crisis levels of aid intensity.125
Setting aside the crisis-related state aid, the amount of
state aid in 2011 was back at the 2007 level. These
very low figures are interesting for a number of
reasons. First of all, they show that the amount of

123
Crisis-related aid measures to the financial sector were subject
to a different set of rules and the amounts involved were much
higher, reaching 1.9% of EU GDP in 2008 and 2.9% of GDP
in 2009. These amounts are not included in Figure 4.9.
124
The measures of the Temporary Framework included the
possibility to grant direct subsidies to individual firms up to an
amount of EUR 500,000; the provision of state guarantees at
reduced premia; additional interest-rate support for loans
financing investments in green products; and the possibility for
official export credit agencies (ECAs) to provide cover for
short-term transactions which were previously considered to be
marketable risk.
125 126
The Temporary Framework expired by the end of December The priority for a fiscal consolidation affected considerably
2011. In the period 2008-2011 about EUR 4.8 billion of state state aid measures in several Member States.
127
aid (0.04% of EU GDP) was paid out under the Temporary Note that so-called de minimis aid provided by Member States
Framework, mainly in the form of subsidies and direct grants is not included in the state aid figures because de minimis aid
(European Commission, 2012b). The Temporary Framework need not be notified to the Commission. De minimis aid
was open to all industries and sectors but de facto the majority represents all aid measures with an aid amount below EUR
of the aid was allocated to car producers which were hit hard 200,000 (this threshold applies since December 2006 when it
by the crisis due to the crisis-related slump in car sales. was raised from EUR 100,000).

118
Table 4.3. Internationalisation measures and competitiveness
Dependent variable: Member States share in total extra-EU exports
Specification (1) (2) (3) (4)

internationalisation measures 0.024 *** 0.020 *** 0.025 *** 0.022 ***
(0.005) (0.006) (0.008) (0.006)
internationalisation measures -0.001 -0.001 0.000 -0.001
(0.001) (0.002) (0.001) (0.001)
loans to GDP 0.071
(0.072)
loans to GDP -0.269 ***
(0.033)
loans to GDP * internationalisation measures -0.009
(0.007)
governance 0.437
(0.356)
governance -0.105
(0.981)
governance * internationalisation measures 0.142 ***
(0.017)
wage share 0.179
(0.422)
wage share 2.224
(1.957)
wage share * internationalisation measures 0.108 **
(0.045)
tariff rate 0.071 *
(0.040)
tariff rate -0.026
(0.030)
tariff rate * internationalisation measures 0.066 ***
(0.012)

R 0.993 0.990 0.989 0.990

adjusted R 0.992 0.988 0.987 0.989
Observations 373 380 341 391
Note: Standard errors appear in parentheses. ***, **, * indicate statistical significance at the 1%; 5% and 10% level respectively.
Regressions include country and year fixed effects as well as a constant term which are not reported. The standard errors are robust. All
the data was logarithmised (observations of the value zero were changed to 0.01 in order to make the taking of logarithms possible) and
centred in order to make the estimated coefficients interpretable.
Source: WIOD, European Union State Aid Scoreboard, Eurostat, UNCTAD-TRAINS, World Banks Worldwide Governance Indicators
(WGI) database.

indicators. The three base specifications (see annex

4.13. QUANTITATIVE ASSESSMENT OF STATE AID
AND EXPORT ORIENTATED MANUFACTURING
Three different specifications have been implemented
with the objective to quantitatively assess in what
way different types of state aid provided by EU
Member States, impact on the export-oriented
manufacturing sector in the EU. The state aid
variables used are explained below in Box 4.2. The
strategy follows recent empirical literature on the
development of the internationally competitive
manufacturing sector, this proxied with three different
for details) deal with the explanation of extra-EU
export shares (following Aghion et al., 2011), value-
added per capita in export-orientated manufacturing
(following Haraguchi and Rezonja, 2011) as well as
real value added growth of export-oriented
manufacturing industries (following Rajan and
Subramanian, 2011). The three approaches are used
in order to cover different aspects of the export-
oriented manufacturing sector (value added export
share, per capita level and real growth). Here we
present selected results of the first two approaches.

119
The full set of results is available in background study In a second attempt the methodology put forward in
of this chapter. Haraguchi and Rezonja (2011) is applied to the
provision of state aid by EU Member States. The aim
Box 4.2. Categories of state aid in the European Union
Non-crisis state aid granted by the Member States to industry and services broadly splits into two types: horizontal
and sectoral.
The concept of horizontal aid, which is aid that is not granted to specific sectors of the economy, derives from the EU
Treaty. It leaves room for the Commission to make policy judgements whereby state aid can be considered
compatible with the internal market if it provides effective support for common policy objectives. Most prominent
here is aid earmarked for research, development and innovation, safeguarding the environment, fostering energy
saving and promoting the use of renewable energy sources; Those categories are followed by regional development,
aid to SMEs, job creation and the promotion of training (European Commission, 2012b).
Research, development and innovation: R&D&I lies at the heart of the Europe 2020 strategy as one of its flagship
initiatives because of its potential to contribute to strengthening the competitiveness of the EU economy and to ensure
sustainable growth, with a target of spending 3% of EU GDP on R&D by 2020.
Environmental protection: State aid here can include aid measures to support energy saving and waste management
or to improve production processes, which have a direct benefit to the environment.
Regional development and cohesion: The aim of regional aid is to develop the economic, social and territorial
cohesion of a Member State and of the EU as a whole. The Commission encourages Member States to grant regional
aid on the basis of multi-sectoral schemes which form part of a national regional policy.
Commerce, export and internationalisation: This is a less used measure that however showed some importance in the
quantitative analysis. It consists of a number of different aid measures such as the promotion of brand image or sales
networks but also officially supported export credits to the extent that they contain an aid element.
State aid earmarked for specific sectors, or sectoral aid includes a number of measures targeting for instance: rescue
and restructuring of firms in difficulty; Sectors covered include shipbuilding; steel; coal; land, sea and air transport;
agriculture; fisheries and aquaculture.
The other positive and significant result in this here is to specify better the relationship by adding
specification is the interaction term between more control variables and to test for the determinants
internationalisation measures and the governance of the single manufacturing industrys importance
effectiveness rank. More internationalisation separately, using a model that tries to explain the real
measures is correlated to even bigger export shares in value-added per-capita of the respective
countries with a higher level of domestic competition manufacturing sector.
(i.e. a higher wage share or, in other words, a smaller
profit share, such as in the Nordic and core EU Explanatory variables are the per capita gross
countries), as can be seen from specification (3). domestic product, population density and natural
Finally, countries with both more internationalisation resource endowment as well as different types of state
measures and more tariff protection have on average aid per capita. The control variables which feature
higher extra-EU export shares (see specification (4)). prominently in the growth literature account for
Ceteris paribus a higher tariff protection might developmental impact on manufacturing while other
support the development of domestic manufacturing variables control for demographic and geographic
capacity and induce additional exports. In fact, the conditions. In order to check the robustness of the
positive coefficient is in line with the classical infant estimated results, additional variables such as the
industry argument (which is based on the existence of private loans to GDP indicator have been included
externalities) if such tariff protection were assumed to but the main results do not change very much.
be temporary.128 The effects of sectoral state aid that Moreover in the regression approach the individual
directly targets the manufacturing sector were also manufacturing industries have been aggregated in two
analysed. In none of the estimated specifications did groups export-oriented industries and industries
the conditional main effect of manufacturing aid focusing on the domestic market, based on an
appear to be significantly different from zero (see exportability measure.
Table A.7 in the annex).

128
It should be noted however, that Member States do not set
tariffs themselves because trade policy is a competence of the
European Commission. Any differences in the tariff rate across
Member States is therefore due to differences in their export
structures.

120
The main findings are reported in Table 4.4. The in countries were the manufacturing sector
level of export-oriented manufacturing value-added development was less dynamic over the past one and
per capita is not affected by sectoral specific a half decades (e.g. the United Kingdom). This sign
manufacturing aid. It is rather a few horizontal aid possibly illustrates a correlation pattern existing
spending items which show signs of correlation, restrictively in the specific sample for this exercise.
Regional aid is found to be positively correlated with
the value-added level. Somewhat surprisingly, risk 4.14. COMPANY-LEVEL ANALYSIS OF
capital aid tends to target economies with lower COMMERCIALISATION PERFORMANCE
levels of per capita export-oriented value added. One
explanation may be that regional aid is more likely to Innovation has been placed at the heart of the Europe
be absorbed by large, internationally operating firms, 2020 agenda as one of the main drivers of economic
while environment and energy saving aid can more growth. In a globalised world, innovative ideas and
easily be absorbed by domestically operating smaller products stimulate exports and sales in general,
firms.129 However, the negative sign found for risk- thereby securing growth and future jobs (Harrison et
al., 2008). As the EU-27 is still behind other major
Table 4.4. State aid and value-added per capita export orientated industries
Dependent variable: manufacturing value added per capita of export industries
Specification (1) (2) (3) (4)

per capita GDP 1.537 *** 1.455 *** 1.343 *** 1.615 ***
(0.206) (0.191) (0.199) (0.182)
per capita GDP 0.443 *** 0.483 *** 0.458 *** 0.491 ***
(0.086) (0.090) (0.083) (0.087)
population density -4.376 *** -4.305 *** -5.095 *** -4.416 ***
(1.026) (1.030) (0.949) (0.981)
resource endowment -0.006 0.008 -0.035 0.003 ***
(0.046) (0.047) (0.041) (0.047)
energy saving aid 0.009
(0.008)
regional aid 0.023 ***
(0.007)
risk capital aid -0.027 ***
(0.005)
training aid 0.008 *
(0.004)

R 0.969 0.969 0.972 0.969

adjusted R 0.964 0.964 0.968 0.964
Observations 286 286 286 286
Note: Standard errors appear in parentheses. ***, **, * indicate statistical significance at the 1%; 5% and 10% level respectively.
Regressions include country and year fixed effects as well as a constant term which are not reported. The standard errors are robust.
Source: European Union State Aid Scoreboard, Eurostat, UN Comtrade

capital aid is not straight forward to be explained. A
possibility might also be that firms in manufacturing-
orientated countries such as Germany or Austria rely
more on banks to finance their needs, and
consequently risk capital and risk capital aid is less
important. In contrast, risk capital is more important
129
As shown in the background study, regional aid and risk
capital aid has no impact on value added for domestically
oriented firms while coefficients for energy saving aid and
training aid are significantly positive.
economies when looking at simple innovation
indicators such as overall R&D expenditures, the
impact of innovation policies on firms innovative
behaviour has been a major concern of policy-makers.
Instruments to address shortage of funding for firms
still differ greatly in the EU. Venture capitalists are
more active in Scandinavian and Anglo-Saxon
countries and public funding is on average more
pronounced in EU-15 countries compared to the
EU-12 countries. When looking at the different

121
settings, an essential question that arises is about the
effectiveness of public innovation support. In this
section, the effects of public innovation support on
the commercialisation of R&D effort will be
evaluated.
To that end, the effect of public funding on private
R&D intensity and innovation output is estimated,
using data from the Community Innovation Survey
(CIS)130. Focusing on the commercialisation of R&D
efforts, innovation output will be measured in terms
of innovative sales.
The EU is usually perceived as less effective at
bringing research to the market compared to its main
competitors such as the US, Japan, and South Korea.
The relative underperformance in research
commercialisation in the EU has been attributed to a
number of factors including the absence of an
entrepreneurial culture and a less developed venture
capital sector131. The discussion about the main
factors explaining the European innovation gap dates
back to the Dosi et al. (2006) much-cited criticism on
the concept of the European paradox, a widely
accepted opinion that Europe does not lag behind the
US in terms of scientific excellence, but lacks the
entrepreneurial capacity of the US to effectively
commercialise inventions and step thereby on an
innovation-driven growth path. In the literature there
are a number of publications that investigate whether
Europes weak commercialization performance can
explain the paradox and whether other explanatory
factors are identified (examples of recent overviews
can be found in Conti and Gaule (2011) and Carlsson
et al. (2009)).
This section focuses on which specific innovation-
related factors and which types of public funding can
be identified as relevant for commercialization
performance. By doing so, a firm-level analysis is
provided with a particular focus on the
commercialisation of R&D efforts. Among the
activities examined are the actual R&D performed
internally and/or acquired from external sources, the
research collaboration activities with different players
(such as customers, suppliers, public research
institutions and other firms), and the firms use of
particular types of public funding for innovation.
130
Following to the Community Innovation Survey, public
funding or public innovation support is defined as credits or
deductions, grants, subsidised loans, and loan guarantees for
innovative activities. The support may come from three
authorities: the EU, national governments and regional
authorities.
131
In this report research commercialisation is defined as a sub-
set of the innovation trajectory. Similarly, the
commercialisation gap is understood as a sub-part of the
innovation gap, focusing here on the latest stages of the
innovation trajectory.
122
The analysis of the market uptake of innovation at the
firm-level is based on the Community Innovation
Survey (CIS) micro-data. The CIS survey is a
European-wide, harmonized data collection on
innovation according to the guidelines of the Oslo
Manual (Eurostat/OECD, 2005). The CIS survey is
organized bi-annually and collects information on
enterprises innovation activities. For consistency and
timing reasons, the anonymized CIS2008 and
CIS2006 surveys were used which cover the time
span 2006-2008 and 2004-2006 respectively.
The CIS2006 covers three countries from the EU-15,
namely Greece, Spain and Portugal and 9 countries
from the EU-12. The CIS2008 covers 15 countries.
Germany, Spain, Italy, Ireland and Portugal are EU-
15 countries. The other 10 countries represented in
the CIS2008 are EU-12 countries which joined the
EU after 2004. Additionally to differences in country
coverage the two waves also differ with respect to
sectoral classifications as the CIS2006 is based on the
NACE Rev. 1.1 whereas the CIS2008 on the NACE
Rev. 2 classification. Results reported below
distinguish broader industry aggregates by broad
technology intensity when considering manufacturing
firms only.132
As a key variable for assessing a firms
commercialization performance, this analysis uses the
firms answers on the innovation commercialization
question in the CIS survey; firms were asked about
the percentage of total turnover from new or
significantly improved goods and services introduced
that were new to your market.
The theoretical rationale of the contextual variable
choices comes from to firm-level innovation analysis
literature which also uses the CIS data. The
importance of collaborative R&D for innovation and
general firm performance is illustrated by among
others Laursen and Salter (2006) and Belderbos et al.
(2004). Cassiman and Veugelers (2006) confirm the
importance of innovation activities. More
specifically, they show that internal knowledge
production as well as external knowledge acquisition
is important for a firm to introduce products new to
the market. The combination of internal and external
R&D activities is often referred to in the literature as
absorptive capacity (Cohen and Levinthal (1990)).
132
The industry class dummies have been defined based on the
post-anonymization NACE codes. These codes differ between
the CIS2006 (derived from NACE Rev. 1) and CIS2008
(derived from NACE Rev. 2) waves of the survey. For the
CIS2006 data the following industry class definitions were
used: low-tech (nace_pro codes DA, DB, DC, 20_21, 22, DN),
medium-tech (nace_pro DF_DG, DH, DI, 27, 28), high-tech
(nace_pro DK, DL, DM). For the CIS2008 wave: low-tech
(nace_pro C10_C12, C13_C15, C16_C18, C31_C33),
medium-tech (nace_pro C19_C23, C24_C25), high-tech
(nace_pro C26_C30).
The literature concerning the introduction of new
products into the market illustrates that it is important
to take into account the firms marketing efforts
(market introduction activities) in the
commercialisation model. As such, public funding
plays a role not only in stimulating the R&D efforts
of firms, but also in promoting successful
commercialisation of research results (Griffith et al.
(2006).
model the R&D expenditures is the dependent
variable which is influenced by its own set of factors
(R&D equation)133. Finally, the commercialisation
performance - expressed as turnover from products
new to the market - is estimated on a set of factors
including (estimated) R&D efforts measure. Thus, the
commercialisation and the R&D equations are
connected as the estimated values of the (log of)
R&D expenditure from the latter serve as the input in
the former equation. This procedure avoids the

Table 4.5. Results of the commercialisation output regressions with CIS2006

Medium Medium-
All manuf. Medium-
All firms EU-15 EU-12 Small firms and large Low-tech high and
firms low tech
firms high tech
Commercialisation performance equation
extramural
0.278*** 0.285*** 0.261*** 0.431*** 0.166*** 0.231*** 0.252*** 0.196*** 0.242***
R&D
(28.26) (20.59) (18.39) (9.30) (7.22) (19.01) (10.70) (9.67) (12.50)
log R&D
0.0521*** 0.0446*** 0.0615*** -0.279 0.170*** 0.0874*** 0.176*** 0.0788*** 0.0151
expenditures
(6.80) (3.41) (6.58) (-1.74) (3.53) (8.67) (8.32) (4.67) (0.96)
log 2004
0.00578*** 0.00650*** 0.00480** 0.00247 0.0229*** 0.00545*** 0.00689** 0.00257 0.00652**
sales
(5.40) (4.27) (3.16) (1.86) (9.69) (4.17) (2.84) (1.25) (2.86)
vertical
0.316*** 0.223*** 0.389*** 0.549*** 0.182*** 0.279*** 0.314*** 0.231*** 0.273***
cooperation
(27.63) (13.13) (24.78) (7.60) (7.83) (20.28) (11.93) (10.21) (12.21)
horizontal
0.0640*** 0.0822*** 0.0165 0.193*** -0.00319 0.0273 0.0206 0.0296 0.0335
cooperation
(4.62) (3.42) (0.98) (4.37) (-0.15) (1.56) (0.59) (1.02) (1.23)
intra-group
0.0162 0.0600** -0.0173 0.173*** -0.0287 0.0421* -0.0901* 0.0146 -0.0328
cooperation
(1.15) (2.61) (-1.00) (3.35) (-1.22) (2.48) (-2.37) (0.53) (-1.17)
local public
0.178*** 0.199*** 0.181*** 0.384*** 0.0678** 0.150*** 0.206*** 0.137*** 0.124***
funding
(13.90) (13.14) (4.99) (4.98) (3.04) (9.78) (7.10) (5.62) (4.84)
national
public 0.108*** 0.160*** 0.038 0.506** -0.0183 0.0421* 0.0148 -0.014 0.146***
funding
(7.78) (7.91) (1.91) (2.91) (-0.43) (2.45) (0.44) (-0.49) (5.29)
EU funding 0.0982*** 0.0426 0.139*** 0.295*** -0.00892 0.0741*** 0.0445 0.118*** 0.048
(5.91) (1.53) (6.80) (3.84) (-0.23) (3.58) (1.17) (3.44) (1.38)
F-test for
models 11800.6*** 5333.3*** 6695.5*** 6288.1*** 5176.9*** 5353*** 2039.0*** 1439.3*** 1401.1***
significance

observations 85238 38127 47111 53915 31323 43897 21677 13080 9140
Note: t statistics appear in parentheses. ***, **, *indicate statistical significance at the 0,1%; 1% and 5% level respectively. Regressions
include country and industry fixed effects as well as a constant term which are not reported.
Source: Community Innovation Survey (CIS)

The model on which the analysis is based represents a
slightly modified version of the CDM model
introduced in Crepon et al. (1998). A firms choice to
conduct/report R&D is estimated in the first stage
(R&D selection equation). At the second stage of the
potential and well-known endogeneity problems. The
details of the model applied together with results
from the first and second stages are reported in the
Annex.

133
Ideally, one should include employment additional to turnover
to control for size of firms. As this analysis is based on the
anonymized CD-ROM version of the CIS data this variable is
not available. Instead we used the two employment size classes
available for all countries: small (<50) and medium and large
(>50) for the estimation of the R&D levels.

123
Table 4.6. Results of the commercialisation output regressions with CIS2008
Medium Medium-
All manuf. Medium-
All firms EU-15 EU-12 Small firms and large Low-tech high and
firms low tech
firms high tech

Commercialisation performance equation

extramural
0.231*** 0.237*** 0.227*** 0.348*** 0.330*** 0.160*** 0.170*** 0.129*** 0.183***
R&D
(25.31) (19.65) (15.99) (24.76) (33.78) (13.13) (7.37) (6.59) (8.89)
log R&D
0.0416*** 0.0262* 0.0526*** -0.00596 -0.237*** 0.0796*** 0.147*** 0.0699*** 0.00911
expenditures
(6.10) (2.47) (5.99) (-0.91) (-44.98) (8.64) (7.73) (5.37) (0.52)
log 2006
0.00537*** 0.00364*** 0.00972*** 0.00457*** 0.00967*** 0.00579*** 0.00653** 0.00503* 0.00559*
sales
(5.85) (3.35) (5.60) (3.80) (5.46) (4.75) (3.09) (2.55) (2.44)
vertical
0.260*** 0.218*** 0.298*** 0.376*** 0.333*** 0.227*** 0.219*** 0.223*** 0.209***
cooperation
(26.18) (16.62) (19.38) (24.17) (29.57) (18.48) (8.67) (11.61) (9.62)
horizontal
0.0742*** 0.0606*** 0.0664*** 0.0946*** 0.0865*** 0.0671*** 0.0853** 0.0615* 0.0645*
cooperation
(6.40) (3.60) (4.09) (4.62) (6.71) (4.35) (3.01) (2.43) (2.42)
intra-group
0.0388*** 0.0514** 0.0359* 0.141*** 0.170*** -0.0311* -0.0522 0.00386 0.0125
cooperation
(3.32) (2.90) (2.33) (6.45) (14.57) (-2.02) (-1.77) (0.16) (0.47)
local public
0.156*** 0.167*** 0.266*** 0.205*** 0.157*** 0.120*** 0.200*** 0.0982*** 0.0868***
funding
(14.87) (14.43) (8.12) (12.65) (11.68) (9.19) (8.45) (4.55) (3.83)
national
public 0.138*** 0.172*** 0.105*** 0.243*** 0.383*** 0.0665*** -0.0368 0.0850*** 0.158***
funding
(11.57) (10.67) (5.58) (14.12) (31.76) (4.40) (-1.21) (3.57) (6.25)
EU funding 0.0815*** -0.0478* 0.167*** 0.174*** 0.258*** 0.0831*** 0.104** 0.034 0.105**
(5.64) (-2.02) (9.26) (6.84) (17.11) (4.65) (3.27) (1.13) (3.25)
F-test for
models 14348.6*** 7318.3*** 7075.4*** 7667.7*** 6242.8*** 6135.6*** 2527.8*** 1678.8*** 1187.3***
significance

observations 98070 48472 49598 60780 37290 47144 23615 15096 8433
Note: t statistics appear in parentheses. ***, **, *indicate statistical significance at the 0,1%; 1% and 5% level respectively. Regressions
include country and industry fixed effects as well as a constant term which are not reported.
Source: Community Innovation Survey (CIS)

The variables used for the analysis have been chosen
in a way to ensure the highest possible compatibility
in definitions between the CIS2006 and CIS2008.
One nonetheless has to be cautious when comparing
the models results from different waves of the CIS
survey. Most of the conclusions below are therefore
formulated based on the coefficients signs and their
statistical significance rather than their size.
In addition, to triangulate the obtained results and
check the models robustness, separate analyses were
performed on different types of firms according to
size (small and large)134, geographic location (EU-15
and EU-12 countries) and for manufacturing firms
according to their production technology intensity
(low-tech, medium-low tech, and medium-high and
high tech as described above).
134
Firms with less than 50 employees are classified as small and
above 50 as medium and large. Due to differences in the size
classification among different countries in CIS only these two
classes can be consistently defined.
The results from the first CDM equation (the R&D
intensity estimations presented in Table A8 and Table
A9 in the Annex) give us particular insights into the
factors which influence firms innovation efforts. The
patterns are rather consistent, and show that acquiring
the R&D services externally and benefitting from
national and EU public funding stand out as
consistent factors positively influencing the firms
R&D activities in two waves of CIS and across
different firm types and classes. R&D collaborations
with suppliers and customers as well as inside the
enterprise group are also found to be important
determinants of the firms R&D efforts.
Concerning the specific question on determinants of
R&D commercialisation at the firm level (see Table
4.5 and Table 4.6), the relationships between the
commercialization performance expressed in terms of
the share of turnover from products and services new
to the market and the main characteristics of the
firms innovation activities present several distinctive
patterns.

124
First, the impact of the R&D efforts on
commercialization is positive and statistically
significant when looking at all firms in the sample as
well as manufacturing firms only. Investigating
different subgroups, a positive, significant effect is
observed in low- and medium-low tech
manufacturing industries, and also when looking at
firms from EU-15 and EU-12 in general. The
relationship between the R&D input and the
commercialisation results is mixed when looking at
the performance of firms divided in groups by their
size. The data from both 2006 and 2008 CIS waves
show that for the small firms the relationship between
the R&D expenditures and the share of turnover new
to the market is not evident. For the large and
medium enterprises CIS2006 indicates the
statistically significant positive relationship between
the R&D input and commercialisation, while in
CIS2008 the opposite picture is observed.
It is observed that the firms which, in addition to their
own R&D, also acquire R&D services externally tend
to have higher share of turnover from innovative
products. This external acquisition of R&D results
can take place as a pure purchase of services, but also
can be acquired in the framework of the inter-firm
R&D cooperation.
Concerning the different forms of R&D cooperation
activities the results are mixed across different groups
and classes of firms. It can be seen that vertical
cooperation (i.e. R&D cooperation with suppliers
and/or customers) is positively associated with higher
commercialization performance in firms coming from
different size classes and different technology
intensity groups.
The importance of customers and suppliers for the
firms innovation is further underlined by the finding
that the R&D cooperation inside the group does not
find broad and consistent support by the regression
results. The R&D collaboration with other companies
occupying the similar position in the value chain
(horizontal cooperation) has been shown as a relevant
positive factor by the CIS2008 data and partially by
the CIS2006.
Finally, the effects of public funding on the
commercialization performance of firms appear to be
positive in most classes and groups of firms
considered. The relationship between the use of local
public R&D support and the commercialisation
performance shows positive and statistically
significant for both CIS waves and across all different
technology intensity domains. The public R&D
support at the national level is positively related to the
share of innovative turnover in 2008 with results
being somewhat more mixed for 2006. According to
CIS2008, the firms appear to have higher
commercialisation performance when making use of
the EU-level public R&D support with the exceptions
125
of the EU-15 and medium-tech subsamples. Results
when using CIS2006 are more mixed though when
significant these are always positive.
Across both CIS waves, a consistently strong and
positive effect of public funding is found especially
for firms in medium-high and high-tech industries
and to a lesser degree for lower tech manufacturing
firms. Regarding firm size, the results using CIS 2006
indicate a stronger effect of public funding on small
firms.
Additional to the direct effect of public funding on
the commercialization performance, there is also an
indirect effect via the increase in R&D. As shown in
the second stage, public funding positively affects
R&D levels. In most cases, the increased R&D effort
is estimated to have a positive effect on the
commercialization performance.
Bringing the most important findings of the above
analyses together allows one to formulate a number
of conclusions regarding the general patterns of
innovation and commercialisation performance of
European firms. At the micro-economic level, when
observing the behaviour of individual firms, the link
between the R&D effort and the commercialisation
performance is rather pronounced and a positive
relationship has been observed in most cases.
But not only the R&D itself, also its origin and the
patterns of R&D cooperation among firms play a role.
It has been observed that acquiring results of external
R&D and vertical cooperation with customers and
suppliers is positively related to the firms market
uptake performance.
The above results provide especially pronounced
evidence of the positive effect of the public R&D
support at different levels. When looking at all firms
as well as manufacturing firms only, a positive and
significant effect of all types of R&D support on
R&D levels as well as commercialisation
performance is observed.
The analysis is also performed for a number of
subsamples, which in turn exhibit some specific
patterns. The results suggest that local R&D support
does positively affect firm commercialisation
performance in all technology intensity and size
classes. The effects of national and EU funding are
positive and significant for all firms and
manufacturing firms only, but mixed results are found
for smaller subsamples. Overall, public funding has
consistently positive effects on innovative sales for
medium-high and high-tech sectors firms, while this
statement is true to a lesser extent for firms in lower
tech industries.
4.15. EFFECTS OF PUBLIC FUNDING FOR FIRMS'
R&D
The previous results suggest there is generally a
positive effect from public funding on R&D levels
and commercialisation performance. Nevertheless, a
major problem that the analysis faces is the
possibility of selection bias. Neither the fact that a
firm applies for funding nor the fact that it receives
public support can be considered random. Firms
receiving public support are, for example, more often
exporting firms, which are likely to be more
productive as well. Moreover, firms in higher-tech
industries and those participating in joint R&D
projects are more often supported, as are firms which
are larger in terms of turnover. Thus, selection clearly
has to be taken into account to be able to produce
credible results.
A more detailed look at geographic aspects reveals,
that the R&D intensity as well as the patent
application propensity of EU-15 firms is well above
that of EU-12 firms. The difference in the patent
application propensity is not a function of firm size
distributions as firms in the matched sample are on
average larger in the EU-12 and thus should have a
higher patent application propensity. However, public
funding has had a significantly positive effect in both
country groups. The effects are quite different for the
other innovation output measure the share of
innovative sales. Overall, this share is found to be
larger in the EU-12 due to faster product upgrading,
but the results indicate no effect of public funding on
the commercialisation phase in this region. This
finding is also rather stable over time when looking at
different measurement waves (CIS4 and CIS5).

Box 4.3. The four-step matching procedure

1. Restriction of the sample to the innovative firms of interest: either all innovative firms, or a subsample of
firms with respect to size, country or industry affiliation
2. Estimation of probability of a firm to receive public funding depending on the following observable
characteristics: size based on employment and turnover, country and industry affiliation, exporter status, a
dummy for multinationals and domestic enterprise groups as well as information on R&D cooperation and
preconditions for R&D (estimated at a previous stage)
3. Matching of firms receiving public support with firms which have a similar probability of getting public
funds but do not receive them. Firms are only matched with other firms in the same country and
employment size class (small: less than 50 employees, medium: between 50 and 250, large: more than
250). Firms with no similar counterpart are excluded from the sample using a threshold for the maximum
allowed difference.
4. The average treatment effect can now be calculated as the mean difference of the matched samples.
In the analysis, matching techniques are applied to Interesting results also emerge from the investigation
check for selection bias. According to a number of of effects along the dimension of firm size. Very
observable characteristics, each firm which receives pronounced effects of public support on R&D input
public support is matched with a firm that does not. as well as output can be found for small firms and
The two groups the treatment group, those firms medium-sized enterprises. SMEs often lack sufficient
receiving public support, and the control group internal funds. Support is vital for them to become
should then be similar according to the considered strong entrants in a competitive market able to fill
observable characteristics. world market niches and deliver innovative products.
Effects on patent application rates are especially
One can then estimate the treatment effect on firms pronounced for larger firms. At the same time, no
that receive public support. The complete procedure significant effect of public support on the share of
is an extended version of the one found in Czarnitzki innovative sales can be found for large firms. One
and Lopes-Bento (2013) and is explained in Box 4.3. reason for this finding is that large firms often split
The results shown in Table 4.7 to Table 4.9 indicate research and production facilities geographically and
that for the full sample, public funding has thus output affects may be generated in other
considerable effects on the R&D input as well as subsidiaries.
output. The average R&D intensity in the treatment
group is 1.6% higher than in the control group (Table The most striking results were obtained with respect
4.7). The probability of firms to apply for a patent to the industry affiliation of firms. On the one hand,
(patent application propensity) increases by 8.4% the analysis shows that innovation projects in
with public funding (Table 4.8) and the share of
innovative sales are on average 3.1% higher for firms
that received public funding (Table 4.9).

126
Table 4.7. R&D intensity
R&D intensity Treated Control Difference T-stat
All firms 0.033 0.017 0.016 13.46 ***
EU-15 firms 0.035 0.018 0.017 13.23 ***
EU-12 firms (CIS4) 0.024 0.013 0.011 3.81 ***
EU-12 firms (CIS5) 0.024 0.012 0.013 4.48 ***
Small 0.041 0.019 0.022 10.25 ***
Medium 0.027 0.014 0.014 7.69 ***
Large 0.029 0.019 0.010 4.66 ***
High-tech 0.069 0.036 0.033 6.27 ***
Medium-high-tech 0.041 0.025 0.016 5.97 ***
Medium-low-tech 0.019 0.011 0.009 4.41 ***
Low-tech 0.020 0.013 0.007 3.22 ***
Food processing 0.015 0.006 0.008 2.23 **
Note: The stratified sample overall contains all CIS4 EU-27 countries; the number of treated firms in each sample is: full sample: 5152,
EU-15: 4338, EU-12: 814 (CIS4), 954 (CIS5), Small: 2090, Medium: 1827, Large: 1235, Domestic enterprise groups: 1580, Foreign
enterprise groups: 411, High-tech 633 firms, Medium-high-tech: 1447, Medium-low-tech: 1131, Low-tech: 902, Food processing: 441
***, ** and * denote tests being significant at a 1, 5 and 10% level, respectively.
Source: Community Innovation Survey (CIS), waves 4 and 5, wiiw estimations

Table 4.8. Patent application propensity

Patent application propensity T-stat
Treated Control Difference
All firms 0.303 0.219 0.084 7.54 ***
EU-15 firms 0.323 0.234 0.089 7.03 ***
EU-12 firms (CIS4) 0.192 0.138 0.054 2.62 ***
EU-12 firms (CIS5) 0.158 0.108 0.050 3.00 ***
Small 0.193 0.128 0.066 4.59 ***
Medium 0.284 0.201 0.082 4.62 ***
Large 0.516 0.399 0.117 4.09 ***
High-tech 0.404 0.288 0.117 3.38 ***
Medium-high-tech 0.435 0.317 0.117 5.08 ***
Medium-low-tech 0.249 0.195 0.055 2.55 **
Low-tech 0.121 0.127 -0.007 -0.35
Food processing 0.163 0.091 0.073 2.51 **
Note: The stratified sample overall contains all CIS4 EU-27 countries; the number of treated firms in each sample is: full sample: 5152,
EU-15: 4338, EU-12: 814 (CIS4), 954 (CIS5), Small: 2090, Medium: 1827, Large: 1235, Domestic enterprise groups: 1580, Foreign
enterprise groups: 411, High-tech 633 firms, Medium-high-tech: 1447, Medium-low-tech: 1131, Low-tech: 902, Food processing: 441.
***, ** and * denote tests being significant at a 1, 5 and 10% level, respectively.
Source: Community Innovation Survey (CIS), waves 4 and 5, wiiw estimations.

higher-tech industries (which basically comprise
advanced manufacturing industries, see Annex Table
A.2) benefit particularly from public funding. This
can be seen from the significant and large effects on
both the patent application propensity and the share
of innovative sales.
Publicly funded firms in high- and medium-high-tech
industries exhibit a higher increase in the share of
innovative sales of 8.7 and 4.1 percentage points,
respectively and an 11.7 percentage points higher
application rate for patents.
On the other hand, the results indicate strong
crowding out effects of public funding in lower-tech
industries, especially with respect to innovation
output measures. The finding is not an effect of
lower-tech EU-12 firms, which overall exhibit no
significant effects of public funds on the share of
innovative sales, but can be found for lower-tech EU-
15 firms as well. A possible explanation is that
innovation projects in these industries take place in an
environment which is changing less rapidly than that
of high-tech industries. Thus, there is on average less
risk and asymmetric information attached to

127
Table 4.9. Share of innovative sales
Share of innovative sales T-stat
Treated Control Difference
All firms 0.232 0.201 0.031 4.11 ***
EU-15 firms 0.222 0.188 0.033 4.04 ***
EU-12 firms (CIS4) 0.288 0.269 0.019 1.09
EU-12 firms (CIS5) 0.285 0.277 0.009 0.57
Small 0.225 0.198 0.027 2.19 **
Medium 0.233 0.190 0.042 3.55 ***
Large 0.244 0.222 0.022 1.37
High-tech 0.336 0.249 0.087 3.84 ***
Medium-high-tech 0.261 0.220 0.041 3.04 ***
Medium-low-tech 0.178 0.166 0.012 0.83
Low-tech 0.200 0.190 0.010 0.60
Food processing 0.173 0.149 0.024 0.92
Note: The stratified sample overall contains all CIS4 EU-27 countries; the number of treated firms in each sample is: full sample: 5152,
EU-15: 4338, EU-12: 814 (CIS4), 954 (CIS5), Small: 2090, Medium: 1827, Large: 1235, Domestic enterprise groups: 1580, Foreign
enterprise groups: 411, High-tech 633 firms, Medium-high-tech: 1447, Medium-low-tech: 1131, Low-tech: 902, Food processing: 441.
***, ** and * denote tests being significant at a 1, 5 and 10% level, respectively.
Source: Community Innovation Survey (CIS), waves 4 and 5, wiiw estimations.

innovation projects in low-tech industries. Banks and
other financial intermediaries can therefore better
evaluate them. Innovation market failures can be
expected to be less pronounced in traditional
industries meaning here is also less need for public
funding. This is especially true for larger firms, which
can either rely on internal funding or have easier
access to external sources such as banks. The finding
also indicates that the increased innovation support
via the Rural Development Policy, which is part of
the European Common Agricultural Policy, has no or
very small effects on innovation output.135 It might
thus be more desirable to reallocate these innovation
funds to a broader support of competitiveness, as is
planned in the budget for the period 2014-2020.
4.16. SUMMARY AND POLICY IMPLICATIONS
Despite of longer-term trends in advanced economies
whereby the manufacturing sector accounts for a
shrinking share of value-added of employment, there
is a considerable case for preserving a critical size
of manufacturing activities in European economies.
The main arguments are the following: Firstly,
manufacturing still accounts for a major part of
innovation effort in advanced economies which
translates into above-average contributions to overall
productivity growth and thus to real income growth.
Secondly, there are very important backward
linkages from manufacturing to services which
provide important inputs for manufacturing (in
135
Food processing was analysed separately, as firms in this
industry exhibit by far the highest support rate with respect to
EU funds.
particular business services). Thirdly, manufacturing
has a carrier function for services which might
otherwise be considered to have limited tradability. In
the same direction goes the increased product
bundling of production and service activities in
advanced manufacturing activities. This carrier
function through international competitive pressure
has furthermore a stimulus effect for innovation and
qualitative upgrading for service activities. Lastly,
and related to the first argument, is the higher
productivity growth in manufacturing which is
important because the sector of origin of productivity
growth be the sector that benefits most from the
actual productivity gain.
The main findings of the analyses of state aid and
export-oriented manufacturing are the following.
Regarding extra-EU manufacturing export shares of
Member States, internationalisation measures appear
to be a support item that has a positive effect. Also, in
the case of internationalisation measures, there is a
positive interaction effect with governance
effectiveness.
As for the per capita levels of the export-oriented
manufacturing sector in the EU-27 countries, regional
aid and training aid were positively associated with
value-added per capita growth while risk capital aid
has a significant negative effect.
The results from the analyses of public funding of
R&D&I suggest that it can be better targeted in the
EU-12 and make it more effective. Especially in the
EU-12, and irrespective of the actual objectives of the
support programmes, de facto governments end up
providing innovation support more often to larger
firms than to their smaller competitors (for detailed

128
evidence refer to the background study of the
report).Given the substantial evidence that small
firms in particular face considerable financial
problems due to asymmetric information impacts,
they should be the primary target of public funds.
In order to increase support to small firms, a special
targeting of grants is one way to improve the
allocation of public funds. Other initiatives could
include information campaigns about credits,
deductions and subsidised loans for new
entrepreneurs. As problems lie mainly in the
commercialisation phase, fostering venture capital
investment would be another starting point.
Industrial policy is designed to improve the growth
process (in its quantitative and qualitative aspects)
through its impact upon economic structures (see also
Pack and Saggi, 2006). This could be done by
impacting economic structure in terms of the
composition of activities or industries, or by
influencing the directions in which technologies
develop or within industries, by affecting the
distribution of enterprises and plants according to
different performance characteristics. There is also
the influence on the distribution of economic activity
over geographic area, so that industrial policy has an
interface with regional policy. The impact of
industrial policy on economic activity may take place
directly (e.g. through direct support for particular
types of industries, firms, technologies) or indirectly
(through framework conditions such as the way
financial markets operate or the legal and
administrative system or the quality of educational
and training institutions).
The second goal of industrial policy apart growth
is external competitiveness, which means that would
pay particular attention to the development of the
tradable sector (in all the dimensions cited previously:
composition of activities and industries; intra-industry
composition; technologies and product quality).
Furthermore, industrial policy has to be attentive to
the different needs of countries and regions at
different levels of economic development.
The maintenance of a competitive and diversified
industrial base is part of the Europe 2020 strategy.
The policy challenge could be seen as providing the
right framework conditions and public inputs so that
gaps do not open up in the spectrum of industrial
activities which could be deemed strategic in terms of
the future development of industrial activity.
Strategic in this context means that such segments
of industrial activity do or could (in future) exert
important spillover effects in terms of backward or
forward learning processes in linked activities and/or
could also provide important inputs for various
activities.
129
Industrial policy at the EU level should ensure that
Europe has a broad and diversified industrial structure
which is well-equipped to be a major actor in the
development of new areas of activity such as
environmental technology. In this it is able to benefit
from the diversified character of European industrial
and demand structures and benefit from the pooling
of resources. This encourages innovations in existing
areas, in which Europe draws on its specific
comparative advantages, to be based on traditions of
production specialisation (fashion in France and Italy,
high-quality mechanical engineering and transport
equipment in Germany and in a number of the central
European economies), or on a diversified pattern of
private and public demand. The latter includes
features of the European model such as the strong
position of public transport, of high-quality health
services or linked medical devices and
pharmaceuticals.
Furthermore, the preservation of the industrial
commons includes nurturing manufacturing-services
inter-linkages and exploiting specialisation
advantages of different European economies. State
aid measures to support structural change and
structural adjustment have so far been used
predominantly at national level and did not rely much
on the coordinated use of state aid tools. In a highly
integrated European economy, the preservation and
development of industrial commons should be seen
as a joint responsibility because of strong externalities
across the European economy. Such joint
responsibility for industrial commons includes rules
for quality assurance and recognition of
qualifications, supporting the mobility of skilled staff,
learning from successful cluster policies, support for
necessary transport and communications
infrastructure.
It is important that concept of industrial policy
support the structural change enhancing rather than
the structure preserving aspects. Industrial policy
should play an active role in reducing entry barriers in
four directions: supporting new firms, developing and
marketing new products, moving into new markets or
market niches.
Something well worth elaborating and increasing is
demand-side industrial policy as an instrument to
stimulate the commercial application of innovation.
There is a broad consensus that the existing gap
between European research excellence and the
development of marketable products is a major weak
spot in Member States innovation systems. The US
defence-related public procurement policy may serve
as an example of how to remedy this shortcoming. As
pointed out above, public procurement can provide
the necessary incentive to invest in the development
of marketable products. Given the strong political
commitment of the EU to environmental protection
and the mitigation of climate change, a long-term
industrial policy targeted at the development of
clean products and technologies could well form the
base for a major industrial policy initiative.
Importantly, such a strategy should not only include a
long-term funding commitment for research but also
needs a reliable source of demand that should be
provided by public procurement of EU Member
States and the EU itself.
The industrial policy strategy laid out in the European
Commissions Industrial Policy Communication of
October 2012 (European Commission, 2012a) goes in
the same direction: Five of the six priority areas
(priority action lines) defined in this Communication
are related to meeting the challenge of climate change
and the degradation of the environment. It remains to
be seen whether public procurement will have any
role to play in the EUs policy initiatives for
stimulating the commercialisation of innovations and
the development of green and more resource-efficient
products.
This issue has been much researched and forms the
backbone of many policy initiatives (most
prominently the Lisbon Agenda, and subsequently in
the Europe 2020 Agenda). In the face of
technological competition, particularly with the
United States and more recently with a range of Asian
economies, innovation has increasingly become the
focus of industrial policy at EU level.
Analyses in this chapter have contributed to the
evaluation of innovation policy in the way it is
conducted at national and EU levels. They come out
in favour of further efforts towards increased
harmonisation of innovation systems and the use of
innovation policies across the EU Member States.
Attempts at EU level have already been made to
create an internal market for research, supporting
130
the free movement of knowledge, researchers and
technology, with the aim of increasing cooperation,
stimulating competition and achieving a better
allocation of resources and an improved coordination
of national research activities and policies (FREE,
2010). The attempts have been further reinforced
through the Article 179 of the Lisbon Treaty, creating
an unified research area based on the internal market,
in which researchers, scientific knowledge and
technology circulate freely and through which the
Union and its Member States strengthen their
scientific and technological bases, their
competitiveness and their capacity to collectively
address grand challenges (European Commission,
2012c).
However, the empirical analysis conducted also
shows that innovation policies conducted at EU,
national and regional levels partly address different
needs, such as support for large firms vs. SMEs,
national enterprise groups vs. multinationals,
activities where the technological spillovers are more
local vs. those which are international. It was also
found that there can be different instances of
misallocation of resources in the way programmes are
conceived at EU or national levels. The different
focus is understandable as issues of asymmetric
information and knowledge of spillover effects are
perceived differently at local, national and EU levels.
Hence a clear view of division of tasks and use of
resources at these different levels is important in the
area of innovation policy as in many other areas.
 REFERENCES
Acemoglu, D., Aghion, P., Bursztyn, L., and Hemous, D., (2010), The Environment and Directed Technical
Change, MIT and Harvard Disc. Paper, Mass.
Aghion, P., Bloom, N., Blundell, R., Griffith, R., and Howitt, P., (2005), Competition and Innovation: An
Inverted-U Relationship, Quarterly Journal of Economics, 120(2), pp. 701-728.
Aghion, P., Boulanger, J., and Cohen, E., (2011), Rethinking Industrial Policy, Bruegel Policy Brief, 2011-04,
http://www.bruegel.org/publications/publication-detail/publication/566-rethinking-industrial-policy/
Aghion, P., Dechezlepretre, A., Hemous, D., Martin, R., and Van Reenen, J., (2010), Path Dependence in Clean
Versus Dirty Innovation: Evidence from the Automotive Industry, unpublished manuscript.
Aghion, P., Dewatripont, M., Du, L., Harrison, A., and Legros, P., (2012), Industrial Policy and Competition,
NBER Working Paper No. 18048.
Aiginger, K., (2007), Industrial Policy: A Dying Breed or a Re-emerging Phoenix?, Journal of Industry,
Competition and Trade, 7(3-4), pp. 297-323.
Aiginger, K., and Sieber, S., (2006), The Matrix Approach to Industrial Policy, International Review of Applied
Economics, 20(5), December, pp. 573-603.
Arrow, K.J., (1962), Economic welfare and the allocation of resources for invention, in Nelson, R.R. (ed.), The
Rate and Direction of Inventive Activity: Economic and Social Factors, National Bureau of Economic
Research, Conference Series, Princeton University Press, Princeton, pp. 609625.
Baldwin, R., and Robert-Nicoud, F., (2010), Trade-in-goods and trade-in-tasks: An Integrating Framework,
NBER Working Papers 15882.
Baumol, W.J., (1967), Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis, American
Economic Review, 57(3), pp. 415-426.
Blauberger, M., (2008), From Negative to Positive Integration? European State Aid Control Through Soft and
Hard Law, Max Planck Institute for the Study of Societies Discussion Paper 08 /4, April.
Belderbos, R., Carree, M. and Lokshin, B. (2004). Cooperative R&D and firm performance. Research Policy,
Vol. 33, pp. 1477-1492.
Bock-Schappelwein, J., Hlzl, W., Janger, J., and Reinstaller, A., (2013), Die Rolle von Bildung fr die
wirtschaftlichen Perspektiven sterreichs, WIFO Monthly Report 2/2013, pp. 121-133.
Canepa, A., and Stoneman, P., (2008), Financial constraints to innovation in the UK: evidence from CIS2 and
CIS3, Oxford Economic Papers, 60, pp. 711-730.
Carlsson, B., Acs, Z., Audretsch, D.B., Braunerhjelm, P., (2009), Knowledge creation, entrepreneurship, and
economic growth: a historical review, Industrial and Corporate Change, Vol. 18., No. 6.
Cassiman, B. and Veugelers, R. (2006). In search of complementarity in innovation strategy: Internal R&D and
external knowledge acquisition. Management Science, Vol. 52, No. 1.
Chang, H.-J., (2006), Industrial policy in East Asia lessons for Europe, in EIB (ed.), An industrial policy for
Europe? From concepts to action, EIB papers, 11(2), pp. 106-133.
Ciriaci, D., Hervas, F. (2012), Bridging ideas with markets - The impact of training, marketing and design on
innovation, JRC Policy Brief.
Cohen, E., (2006), Theoretical foundations of industrial policy, EIB Papers, 11(1), pp. 84-106.
Cohen, W. M. and Levinthal, D.A. (1990). Absorptive Capacity: A New Perspective on Learning and
Innovation. Administrative Science Quarterly, Vol. 35, No. 1.
Crafts, N., (2010), Overview and Policy Implications, in BIS (ed.), Learning from some of Britains Successful
Sectors: An Historical Analysis of the Role of Government, BIS Economics Papers, No. 6, March, pp. 1-17.
Conti, A., Gaule, P. (2011), Is the US outperforming Europe in university technology licensing? A new
perspective on the European Paradox, Research Policy, Vol. 40., No. 1.
Crepon, B., Duguet, E., and Mairesse J. (1998), Research, Innovation and Productivity: An Econometric
Analysis at the Firm Level. Economics of Innovation and New Technology, Vol. 7, No. 2.

131
Criscuolo, C., Martin, R., Overman, H., and Van Reenen, J., (2012), The Causal Effects of an Industrial Policy,
IZA Discussion Paper, 6323, Bonn.
Czarnitzki, D., and Lopes-Bento, C., (2013), Value for money? New microeconometric evidence on public
R&D grants in Flanders, Research Policy, 42(1), pp. 76-89.
Dachs, B., Biege, S., Borowiecki, M., Lay, G., Jger, A., and Schartinger, D., (2013), Servitisation in European
manufacturing industries: empirical evidence from a large-scale database, The Service Industries Journal.
Doleys, T.J., (2012), Managing the Dilemma of Discretion: The European Commission and the Development of
EU State Aid Policy, Journal of Industry, Competition and Trade (first online doi 10.1007/s10842-012-
0140-y).
Dosi, G., Llerena, P., Labini, M.S. (2006), The relationships between science, technologies and their industrial
exploitation: An illustration through the myths and realities of the so-called European European Paradox,
Research Policy, Vol. 35., No. 10.
Eaton, J., Gutierrez, E., and Kortum, S., (1998), European technology policy, Economic Policy, 13(27), pp.
403-438.
European Commission, (2006), Emerging Science and Technology Priorities in Public Research Policies in the
EU, the US and Japan, March 2006, http://ec.europa.eu/research/social-sciences/pdf/ntw-emerging-
report_en.pdf.
European Commission, (2010a), An integrated industrial policy for the globalisation era. Putting
Competitiveness and Sustainability at Centre Stage, COM(2010)614, Brussels.
European Commission, (2010b), Europe 2020. A strategy for smart, sustainable and inclusive growth,
COM(2010) 2020 final, Brussels.
European Commission (2011a), European Competitiveness Report 2011, Commission staff working document
SEC(2011) 1188 accompanying document to the Communication Industrial Policy:reinforcing
competitiveness (COM(2011) 642 final) (SEC(2011) 1187 final). Luxemburg.
European Commission, (2010b), Europe 2020. A strategy for smart, sustainable and inclusive growth,
COM(2010) 2020 final, Brussels.
European Commission (2011a), European Competitiveness Report 2011, Commission staff working document
SEC(2011) 1188 accompanying document to the Communication Industrial Policy:reinforcing
competitiveness (COM(2011) 642 final) (SEC(2011) 1187 final). Luxemburg.
European Commission, (2011b), State Aid Scoreboard. Report on state aid granted by the EU Member States,
Autumn 2011 Update, COM(2011)848 final.
European Commission (2011c), Europe 2020 Flagship InitiativeInnovation Union, SEC(2010) 1161.
European Commission (2011d), Innovation Union Scoreboard 2011, PRO-INNO Europe.
European Commission, (2012a), A Stronger European Industry for Growth and Economic Recovery, Industrial
Policy Communication Update, 10 October 2012 COM(2012)582/3.
European Commission, (2012b), Facts and figures on State aid in the EU Member States,. 2012 update,
Commission Staff Working Document accompanying the document State aid Scoreboard 2012 Update
Report on State aid granted by the EU Member States, COM(2012) 778 final.
European Commission, (2012c), A Reinforced European Research Area Partnership for Excellence and
Growth, COM(2012) 392 final, Brussels.
Eurostat/OECD (2005). Oslo-Manual Proposed guidelines for collecting and interpreting technological
innovation data. 3rd ed. (Paris)
Foster, N., Pschl, J., and Stehrer, R. (2012), Offshoring and the Elasticity of Labour Demand, wiiw Working
Papers, No. 90, July.
FREE, (2010), Innovation Systems Guidebook, FREE From Research to Enterprise, April.
Friederiszick, H.W., Rller, L.-H., and Verouden, V., (2006), European State Aid Control: An Economic
Framework, in Buccirossi, P. (ed.), Advances in the Economics of Competition Law, Cambridge, MA, MIT
Press.
Frst, E., (2013), Renaissance der Industriepolitik?, GfE Policy Brief, 05 2013.

132
Garcia, A. (2011), The Relevance of marketing in the success of innovations, JRC Technical notes No. 9.
Goos, M., Manning, A., and Salomons, A., (2010), Explaining Job Polarization in Europe: The Role of
Technology, Globalization and Institutions, CEPS Discussion Paper, No. 1026, November.
Griffith R., Huergo, E., Mairesse J. and Peters B. (2006), Innovation and Productivity across four European
Countries., Oxford Review of Economic Policy, Vol. 22.
Grossman, G., and Rossi-Hansberg, E., (2008), Trading Tasks: A Simple Theory of Offshoring, American
Economic Review, 98(5), pp. 1978-1997.
Hall, B.H., (2002), The Financing of Research and Development, Oxford Review of Economic Policy, 18,
pp. 35-51.
Hall, B.H., and Lerner, J., (2009), The Financing of R&D and Innovation, NBER Working Papers, 15325,
National Bureau of Economic Research, Inc.
Haraguchi, N., and Rezonja, G., (2011), Unravelling Manufacturing Development: The Role of Comparative
Advantage, Productivity Growth and Country-specific Conditions, UNIDO Development Policy, Statistics
and Research Branch Working Paper, No. 16/2011.
Harrison, R., Jaumandreu, J., Mairesse, J., and Peters, B., (2008), Does Innovation stimulate employment? A
firm level analysis using comparable micro-data from four European countries, NBER Working Paper
14216, National Bureau of Economic Research, Inc.
Hausmann, R., and Rodrik, D., (2003), Economic development as self-discovery, Journal of Development
Economics, 72(2), pp. 603-633.
Hausmann, R., and Rodrik, D., (2006), Doomed to choose: Industrial Policy as a Predicament, September 2006
, http://www.hks.harvard.edu/fs/drodrik/Research%20papers/doomed.pdf.
Helper, S., Krueger, T., and Wial, H., (2012), Why Does Manufacturing Matter? Which Manufacturing
Matters? A Policy Framework, Brookings, February.
Hijzen, A., and Swaim, P., (2007), Does offshoring reduce industry employment?, GEP Research Paper
no. 2007/24, The Leverhulme Centre for Research on Globalisation and Economic Policy, The University of
Nottingham.
Hijzen, A., and Swaim, P., (2010), Offshoring, labour market institutions and the elasticity of labour demand,
European Economic Review, 54, pp. 1016-1034.
Himmelberg, C.P., and Petersen, B.C., (1994), R&D and Internal Finance: A Panel Study of Small Firms in
High-Tech Industries, Review of Economics and Statistics, 76, pp. 38-51.
Hippach-Schneider, U., Krause, M., and Woll, C., (2007), Vocational education and training in Germany. Short
description, Cedefop Panorama series, 138.
Johnson, R.C., and Noguera, G., (2012), Accounting for Intermediates: Production Sharing and Trade in Value
Added, Journal of International Economics, 86(2), pp. 224-236.
Kaldor, N., (1968), Productivity and Growth in Manufacturing Industry: A Reply, Economica, 35(140),
pp. 385-391.
Langlois, R.N., and Steinmueller, E., (1999), The Evolution of Competitive Advantage in the Worldwide
Semiconductor Industry, 1947-1996, in Mowery, D.C., and Nelson, R.R. (eds), Sources of Industrial
Leadership. Studies of Seven Industries, Cambridge.
Laursen, K. and Salter, A. (2006), Open for innovation: the role of openness in explaining innovation
performance among U.K. manufacturing firms. Strategic Management Journal, Vol. 27.
Mayer, J., (2004), Industrialization In Developing Countries: Some Evidence From A New Economic
Geography Perspective, UNCTAD Discussion Papers 174, United Nations Conference on Trade and
Development.
McKinsey Global Institute, (2012), Manufacturing the future: The next era of global growth and innovation,
November.
Midelfart, K.H., and Overman, H.G., (2002), Delocation and European integration: is structural spending
justified?, Economic Policy, 35, pp. 322-359.

133
Myers, S.C., and Majluf, N.S., (1984), Corporate Financing and Investment Decision when Firms have
Information that Investors do not have, Journal of Financial Economics, 13, pp. 187-221.
Naud, W., (2010a), Industrial Policy. Old and New Issues, UNU-WIDER Working Paper No. 2010/106,
September.
Naud, W., (2010b), New Challenges for Industrial Policy, UNU-WIDER Working Paper No. 2010/107,
September.
Nords, H.K., and Kim, Y., (2013), The Role of Services for Competitiveness in Manufacturing, OECD Trade
Policy Papers, 148, OECD Publishing.
Norman, V. D., & Orvedal, L., (2010), En liten, pen konomi (4. utg.). Oslo: Gyldendal Akademisk.
Owen, G., (2012), Industrial policy in Europe since the Second World War: What has been learnt?, ECIPE
Occasional Paper no. 1, The European Centre for International Political Economy, Brussels.
Pack, H., and Saggi, K., (2006), Is There a Case for Industrial Policy? A Critical Survey, World Bank Research
Observer, 21(2), pp. 267-297.
Page, A.L., (1993), Assessing new product development practices and performance: Establishing crucial
norms, Journal of Product Innovation Management, 10, pp. 273-287.
Pelkmans, J., (2006), European Industrial Policy, Bruges European Economic Policy Briefings no. 15, College of
Europe, Bruges.
Pisano, G.P., and Shih, W.C., (2009), Restoring American Competitiveness, Harvard Business Review, 87,
July-August, pp. 114-125.
Polder, M., Veldhuizen, E.,van den Bergen, D. and van der Pijll, E. (2009), Micro and macro indicators of
competition: comparison and relation with productivity change, CBS Discussion Paper 09024.
Pollin, R., and Baker, D., (2009), Public Investment, Industrial Policy and U.S. Economic Renewal, Joint CEPR-
PERI Working Paper 211.
Rajan, R.G., and Subramanian, A., (2011), Aid, Dutch disease, and manufacturing growth, Journal of
Development Economics, 94(1), pp. 106-118.
Reiner, C., (2012), Play it again, Sam: die Renaissance der Industriepolitik in der Groen Rezession, Wirtschaft
und Gesellschaft, 38(1), pp. 15-56.
Reinstaller, A., Hlzl, W., Kutsam, J., and Schmid, C., (2012), The development of productive structures of EU
Member States and their international competitiveness, WIFO research study.
Rodrik, D., (2012), The Manufacturing Imperative, project-syndicate, 10 August 2012, http://www.project-
syndicate.org/commentary/the-manufacturing-imperative.
Rubenstein, A.H., Chakrabarti, A.K., OKeefe, R.D., Souder, W.E., and Young, H.C., (1976), Factors
influencing innovation success at the project level, Research Management, 19, pp. 15-19.
Senses, M.Z., (2010), The effects of offshoring on the elasticity of labor demand, Journal of International
Economics, 81, pp. 89-98.
Soete, L., (2009), Matlthus Revenge, in Faroult, E. (ed.), The World in 2025. Contributions from an expert
group, European Commission Working Document.
Spielkamp, A., and Rammer, C., (2009), Financing of innovations thresholds and options, Management &
Marketing, 4, pp. 3-8.
Stehrer, R., (2012), Trade in Value Added and the Value Added in Trade, WIOD Working Paper 8.
Stllinger, R., (2011), Global Patterns of Trade Specialisation and Club Convergence: Past and Current Patterns
of Trade Specialisation, AUGUR Working Paper, September.
Timmer, M. (2012, ed.), The World Input-Output Database (WIOD): Contents, sources and methods, WIOD
Working Paper 10.
UNIDO, (2002), Industrial Development Report 2002/2003. Competing through Innovation and Learning,
UNIDO, Vienna.

134
Van Pottelsberghe, B., (2007), Europes R&D: Missing the wrong Targets?, Bruegel policy brief, issue 2008/3,
http://www.bruegel.org/publications/publication-detail/publication/7-europes-r-and-d-missing-the-wrong-
targets/.
Wade, R.H., (2012), Return of industrial policy?, International Review of Applied Economics, 26(2),
pp. 223-239.
Warwick, K., (2013), Beyond Industrial Policy. Emerging issues and new trends, OECD Science, Technology
and Industry Policy Papers, 2, OECD Publishing.

135
 ANNEX 1
INDUSTRY LISTS AND COUNTRY ABBREVIATIONS

Table A1. Country abbreviations

AT Austria
BE Belgium
BG Bulgaria
CY Cyprus
CZ Czech Republic
DE Germany
DK Denmark
ES Spain
EE Estonia
FI Finland
FR France
UK United Kingdom
EL Greece
HU Hungary
IE Ireland
IT Italy
LT Lithuania
LU Luxembourg
LV Latvia
MT Malta
NL Netherlands
PL Poland
PT Portugal
RO Romania
SK Slovakia
SI Slovenia
SE Sweden
US Unites States
JP Japan
KR South Korea
BR Brazil
CN China
IN India

136
Table A2. Industry classification with detailed advanced manufacturing industries
15t16 Food, Beverages and Tobacco Low technology
17t18 Textiles and Textile Products Low technology
19 Leather, Leather and Footwear Low technology
20 Wood and Products of Wood and Cork Low technology
21t22 Pulp, Paper, Paper , Printing and Publishing Low technology
23 Coke, Refined Petroleum and Nuclear Fuel Medium-low technology
24 Chemicals and Chemical Products Chemicals
25 Rubber and Plastics Medium-low technology
26 Other Non-Metallic Mineral Medium-low technology
27t28 Basic Metals and Fabricated Metal Metals
29 Machinery, nec Machinery
30t33 Electrical and Optical Equipment Electrical equipment
34t35 Transport Equipment Transport equipment
36t37 Manufacturing, nec; Recycling Low technology
Note: Based on NACE Rev. 1 industry classification.

Table A3. Industry classification according to technology intensity

15t16 Food, Beverages and Tobacco Low technology
17t18 Textiles and Textile Products Low technology
19 Leather, Leather and Footwear Low technology
20 Wood and Products of Wood and Cork Low technology
21t22 Pulp, Paper, Paper , Printing and Publishing Low technology
23 Coke, Refined Petroleum and Nuclear Fuel Medium-low technology
24 Chemicals and Chemical Products Medium-high and high technology
25 Rubber and Plastics Medium-low technology
26 Other Non-Metallic Mineral Medium-low technology
27t28 Basic Metals and Fabricated Metal Medium-low technology
29 Machinery, nec Medium-high and high technology
30t33 Electrical and Optical Equipment Medium-high and high technology
34t35 Transport Equipment Medium-high and high technology
36t37 Manufacturing, nec; Recycling Low technology
Note: Based on NACE Rev. 1 industry classification.

137
Table A4. Industry classification according to Eaton et al. (1998)
15t16 Food, Beverages and Tobacco Labour-intensive / Chemical-linked
17t18 Textiles and Textile Products Labour-intensive / Chemical-linked
19 Leather, Leather and Footwear Labour-intensive / Chemical-linked
20 Wood and Products of Wood and Cork Resource-intensive / Earth-linked
21t22 Pulp, Paper, Paper , Printing and Publishing Resource-intensive / Earth-linked
23 Coke, Refined Petroleum and Nuclear Fuel Resource-intensive / Earth-linked
24 Chemicals and Chemical Products Chemicals
25 Rubber and Plastics Labour-intensive / Chemical-linked
26 Other Non-Metallic Mineral Resource-intensive / Earth-linked
27t28 Basic Metals and Fabricated Metal Metals
29 Machinery, nec Machinery
30t33 Electrical and Optical Equipment Electrical equipment
34t35 Transport Equipment Transport equipment
36t37 Manufacturing, nec; Recycling Resource-intensive / Earth-linked
Note: Based on NACE Rev. 1 industry classification.

138
 ANNEX 2
ADDITIONAL RESULTS FROM THE QUANTITATIVE
ANALYSIS OF STATE AID
Table A5. Percentile ranks of EU Member States' governance effectiveness, average 1995-2011
AT 93.4
BE 93.8
BG 56.4
CY 91.5
CZ 81.5
DE 91.9
DK 99.5
ES 82.0
EE 84.8
FI 100
FR 88.2
UK 92.4
EL 66.8
HU 73.0
IE 89.1
IT 66.4
LT 72.0
LU 94.8
LV 72.5
MT 82.9
NL 96.7
PL 71.6
PT 78.7
RO 47.4
SK 76.3
SI 79.6
SE 98.6
Note: Percentile range (globally) is from 0-100. Higher percentiles indicate higher governance effectiveness.
Source: World Banks Worldwide Governance Indicators (WGI) database.

139
Table A6. Aid to research, development and innovation and competitiveness136
Dependent variable: Member States share in total extra-EU exports
Specification (1) (2) (3) (4)

R&D aid -0.004 0.001 0.011 -0.013

(0.018) (0.022) (0.026) (0.020)
R&D aid -0.001 0.000 0.001 -0.004 *
(0.002) (0.002) (0.003) (0.002)
loans to GDP 0.035
(0.069)
loans to GDP -0.253 ***
(0.035)
loans to GDP * R&D aid -0.002
(0.011)
governance 0.454
(0.342)
governance 0.658
(0.964)
governance * R&D aid -0.134 ***
(0.033)
wage share -0.376
(0.401)
wage share 2.853
(2.276)
wage share R&D aid -0.248 ***
(0.061)
tariff rate -0.046
(0.052)
tariff rate -0.081 ***
(0.029)
tariff rate * R&D aid -0.020 *
(0.012)

R 0.992 0.989 0.988 0.988

adjusted R 0.991 0.987 0.987 0.986
Observations 373 380 341 391
Note: Standard errors appear in parentheses. ***, **, *indicate statistical significance at the 1%; 5% and 10% level respectively.
Regressions include country and year fixed effects as well as a constant term which are not reported. The standard errors are robust. All the
data was logarithmised (observations of the value zero were changed to 0.01 in order to make the taking of logarithms possible) and centred
in order to make the estimated coefficients interpretable. R&D aid is aid to research, development and innovation.
Source: WIOD, European Union State Aid Scoreboard, Eurostat, UNCTAD-TRAINS, World Banks Worldwide Governance Indicators
(WGI) database.

136
It is important to note that this analysis is of a general nature and does not imply a specific link between a certain type of aid and the
trade performance of any particular product or sector.

140
Table A7. Sectoral aid to manufacturing and competitiveness137
Dependent variable: Member States share in total extra-EU exports
Specification (1) (2) (3) (4)

manufacturing aid 0.002 -0.008 0.000 -0.004

(0.006) (0.007) (0.007) (0.006)
manufacturing aid -0.001 -0.001 -0.001 -0.001
(0.001) (0.001) (0.001) (0.001)
loans to GDP 0.054
(0.070)
loans to GDP -0.250 ***
(0.031)
Loans * manufacturing aid 0.004
(0.008)
governance 0.446
(0.362)
governance 0.889
(1.020)
governance * manufacturing aid 0.071 ***
(0.023)
wage share -0.091
(0.399)
wage share 1.436
(1.983)
wage share * manufacturing aid -0.091 **
(0.044)
tariff rate 0.007
(0.042)
tariff rate -0.090 ***
(0.032)
tariff rate * manufacturing aid 0.004
(0.007)

R 0.992 0.988 0.988 0.988

adjusted R 0.991 0.987 0.986 0.986
Observations 373 380 341 391
Note: Standard errors appear in parentheses. ***, **, *indicate statistical significance at the 1%; 5% and 10% level respectively.
Regressions include country and year fixed effects as well as a constant term which are not reported. The standard errors are robust. All the
data was logarithmised (observations of the value zero were changed to 0.01 in order to make the taking of logarithms possible) and centred
in order to make the estimated coefficients interpretable. Manufacturing aid is sectoral aid to manufacturing.
Source: WIOD, European Union State Aid Scoreboard, Eurostat, UNCTAD-TRAINS, World Banks Worldwide Governance Indicators
(WGI) database.

137
It is important to note that this analysis is of a general nature and does not imply a specific link between a certain type of aid and the
trade performance of any particular product or sector.

141
 ANNEX 3
METHODOLOGY
Decomposition of manufacturing R&D intensity
The results for the decomposition of R&D intensities in Figure 4.6. are derived followings the approach of Eaton
et al. (1998). The decomposition approach takes the following form:

where denotes R&D intensity in the manufacturing sector and denotes R&D intensity in industry
i. Subscript c denotes countries and subscript w denotes the global average which for this purpose is the average
of Finland, France, Germany, United Kingdom, Austria, Belgium and the Netherlands as well as the United
States and Japan, i.e. the nine countries included in the decomposition exercise. The valued added shares of
manufacturing are denoted by .

Therefore the first term represents the composition effect, i.e. the differences in industry specialisation across
countries and the second term captures the differences in the industry level R&D intensities. The last term is an
interaction term between those two which has no particular economic interpretation.

Calculation of value added exports

The concept of value added exports used throughout this Report is that of Johnson and Noguera (2012). The
value added exports approach requires global input-output data. In this chapter the world input-output database
(WIOD) is used for this purpose. The WIOD contains information on 40 countries plus the rest of the world
(ROW) for 35 industries. The global input-output table in the WIOD that summarises the inter-industry linkages
is therefore of dimension 1435 x 1435.

The starting point for calculating value added exports (VAX) is the basic input-output identity

where denotes a vector of gross output for each country and industry (i.e. of dimension 1435x1), is a matrix
of intermediate inputs per unit of gross output (of dimension 1435x1435) and is a vector of final demand by
country and sector and therefore again of dimension (1435x1). A final product, e.g. a car, is made of many other
parts produced in other industries maybe even in other countries.

The calculation of VAX consists of decomposing the output vector q of each country r in
where denotes the output absorbed in country r that was sourced from partner
country 1 and likewise for the other partner countries. The elements of q are also referred to as output transfers.
These output transfers are in turn used to calculate the value added produced in a source country i and absorbed
in another country r which constitutes the bilateral value-added exports (VAXi,r).

Bilateral value added exports are defined as , where is the ratio of value added to gross
output in country i and qjr is the output produced in country i that is absorbed in r (see Johnson and
Noguera, 2012). The global value-added exports of country r are obtained by summing up the bilateral
value added exports for all partner countries. The market share of each country in global value added exports
used in the text is then simply .

Quantitative analysis of state aid

Section 4.13 uses three types of approaches to estimate the relationship between the provision of state aid by
Member States and export market shares, value added and value added growth respectively. The empirical
approaches are briefly outlined below.

Aghion, Boulanger and Cohen (2011) type equation: In its basic form the following panel data equation is being
estimated:

142
 ,

where lnEXit represents the log of the overall share of extra-EU manufacturing and services exports of an
EU Member State i in the sample to total EU exports in year t. The variable SA covers total sectoral state aid to
industry and services (also all the other types and sub-groups of state aid are being controlled for) and PC is a
proxy for financial development, measured by the ratio of private credit by deposit-taking banks and other
financial intermediaries to GDP (similarly also indicators of governance, competition and tariff protection are
being checked). The squared terms control for non-linearity and the interaction term checks whether the two
explanatory variables are substitutes or complements. Finally, i and t are country and time fixed effects
respectively, while it is the error term and the s are the coefficients to be estimated. The rationale of this
estimation exercise is to find out whether state subsidies can act as a promoter of international competitiveness,
especially in those cases where access to private finance is limited. (It is important to note that this analysis is of
a general nature and does not imply a specific link between a certain type of aid and the trade performance of
any particular product or sector.) While the original sample of Aghion, Boulanger and Cohen (2011) included
EU-15 data for the years 1992-2008, here EU-27 data for the period 1995-2011 are exploited.

Haraguchi and Rezonja (2011) type equation: The following modified base-line equation is being estimated:

where lnVAjit is the log of the real value added per capita of the respective manufacturing sector j in country i and
year t. The variable DP accounts for the per capita gross domestic product, PD stands for population density and
NR is an indicator for natural resource endowment. Following Haraguchi and Rezonja (2011), the modified
natural resource proxy variable can be calculated as the ratio between exports and imports of crude natural
resource commodities. The commodities included are those categorised under SITC Rev. 1 in Code 2 (crude
materials, inedible, except fuels), 32 (coal, coke and briquettes), 331 (petroleum, crude and partly refined) and
3411 (gas, natural).

These three explanatory variables are seen as mostly exogenous for the specific sample analysed. Here, SA is
state aid per capita, and the s, i and t are defined as in the earlier equation. jit is the error term. The value
added data was taken from Eurostats intermediate ISIC aggregation Rev. 2. GDP and population density data
stems also from Eurostat. Data for constructing the natural resource endowment indicator were taken from the
Comtrade database. In the preferred regressions the single manufacturing sectors have been aggregated in two
groups export-oriented industries and industries focusing on the domestic markets, based on an exportability
measure, in order to make the results better interpretable. In following Rajan and Subramanian (2011) the
exportability of an industry is assumed if the respective industry has a ratio of exports to value added that
exceeds the industry median. For each industry, the median ratio of exports to value added was calculated using
data from all EU-27 countries. The industries above the median are manufacturers of petroleum products,
chemicals, pharmaceuticals, electronics, machinery and cars. Those below are manufacturing food, textiles,
paper, plastics, metals, electric and other equipment.

Rajan and Subramanian (2011) type equation: The basic equation estimated is the following:

where the MGij variable depicts the average annual real growth rate of manufacturing value added of industry j in
country i over the period 2000-2010. A country- and industry-specific indicator of the initial manufacturing
share (IS) is added to the regression in order to control for convergence. Most importantly an interaction term of
state aid as a share of GDP (SA) and a manufacturing sector-specific exportability dummy variable (ED) is
included as well. Similarly to the regression before and following Rajan and Subramanian (2011) the
exportability dummy takes a value of 1 if the respective industry has a ratio of exports to value-added that
exceeds the industry median. For each industry, the median ratio of exports to value added was calculated using
data from all the EU-27 countries. The aim of this regression equation is to check in what way public subsidies
influence the growth of the export-oriented manufacturing sectors in Europe. All the data used have the same
origin as in the second approach.

143
Estimation of commercialisation output model specification
The estimation procedure below for the commercialisation output model specification (used in 4.14) follows a so
called CDM-approach (see Crepon et al. (1998) and Griffith et al. (2006) for more detail) towards estimating the
innovation-driven economic performance of firms based on the CIS data. The CDM procedure uses a multiple
equation econometric model estimate the economic outcomes from the firms innovation efforts.

When estimating the R&D intensity equation using Heckman procedure, the firms decision to perform/report
R&D has been considered as depending on such specific factors as: the firms size represented by the logarithm
of total sales ( log( Si ) ) in the previous period, whether or not the firm is a member of a group ( GR i ):

RDperforme ri i 1 log( Si ) 2 GR i it

The first equation has the logarithm of the firms R&D expenditures as dependent variable and is estimated
conditional on the firms decision to perform/report R&D above:

ln( RDi ) i 1exRDi 2VCoopi 3 HCoopi 4 GCoopi ,

5 LPSi 6 NPSi 7 EUSi 8 FSizei i

where the explanatory variables are the following138:

Extramural R&D indicator, exRDi (1/0);

Vertical Cooperation indicator, VCoopi (1/0);
Horizontal Cooperation indicator, HCoopi (1/0);
Cooperation inside the group indicator, GCoopi (1/0);
Local public funding indicator, LPS i (1/0);
National Public funding indicator, NPS i (1/0);
EU funding indicator, EUS i (1/0)
Firm size class, FSizei (0: <50 employees, 1: >=50).
The second equation is estimated by the means of the tobit regression where the dependent variable is the share
of the turnover from products and services new to the market ( Yi ):

ln(Yi ) i 1exRDi 2 ln( RDi ) 3 log( Si ) 4VCoopi 5 HCoopi

6GCoopi 7 LPSi 8 NPSi 9 EUSi it

The additional explanatory variables are the following:

Predicted value of the logarithm of total R&D, ln( RDi ) ;

The logarithm of total sales in the previous period, log( Si ) .

The estimations also take into account the country-specific intercepts and industry class dummies in order to
correct for individual effects.

138
The innovation activities and funding indicators are taking the value one if they engaged in the past three years in some innovation
activities respectively if they received public funding for innovation activities and the value zero if not. The dummy variables for co-
operation partner takes the value one if the firm indicated a certain type of collaboration in their country or other countries in Europe or
the US or China/India or all other countries and the value zero if not.

144
Table A8. Results of the R&D regressions with CIS2006
Medium Medium-
Small All manuf. Medium-
All firms EU-15 EU-12 and large Low-tech high and
firms firms low tech
firms high tech

R&D expenditures equation

extramural
0.367*** 0.266*** 0.504*** 0.275*** 0.414*** 0.365*** 0.375*** 0.455*** 0.253**
R&D

(8.48) (5.88) (6.07) (4.68) (6.65) (6.96) (3.66) (5.58) (2.84)

vertical
0.429*** 0.328*** 0.557*** 0.436*** 0.398*** 0.370*** 0.455*** 0.385*** 0.242*
collaboration

(7.82) (5.52) (5.51) (5.69) (5.17) (5.59) (3.49) (3.71) (2.19)

horizontal
0.276*** 0.349*** 0.176 0.238* 0.306*** 0.215* 0.129 0.409** 0.103
collaboration

(4.08) (4.13) (1.58) (2.40) (3.34) (2.53) (0.72) (3.06) (0.77)

intra-group
0.468*** 0.507*** 0.384** 0.283* 0.372*** 0.541*** 0.626*** 0.449*** 0.579***
collaboration

(6.52) (6.09) (3.12) (2.24) (4.16) (6.22) (3.30) (3.33) (4.20)

local public
0.381*** 0.354*** 0.707** 0.464*** 0.317*** 0.333*** 0.211 0.308** 0.453***
funding

(6.19) (6.69) (2.84) (5.96) (3.32) (4.57) (1.44) (2.73) (3.76)

national
public 0.986*** 0.954*** 1.003*** 1.076*** 0.853*** 1.006*** 0.915*** 1.031*** 1.037***
funding

(18.30) (17.50) (9.13) (14.31) (11.22) (15.82) (6.87) (10.45) (10.24)

EU funding 0.589*** 0.666*** 0.491*** 0.444*** 0.715*** 0.561*** 0.570** 0.466** 0.678***

(7.58) (7.17) (3.77) (3.86) (6.85) (5.67) (2.94) (2.88) (4.26)

firm size class 0.984*** 0.858*** 1.104*** 0.919*** 0.773*** 0.913*** 1.045***

(20.01) (17.01) (11.07) (15.30) (6.59) (9.72) (10.23)

R&D selection equation

log 2004 sales 0.0370*** 0.0230*** 0.0646*** 0.00322 0.111*** 0.0442*** 0.0502*** 0.0374*** 0.0450***

(21.32) (10.68) (21.03) (1.62) (22.91) (19.56) (12.83) (10.09) (10.67)

member of a
0.424*** 0.418*** 0.441*** 0.357*** 0.246*** 0.501*** 0.545*** 0.505*** 0.447***
group

(33.72) (24.65) (23.17) (18.15) (13.37) (29.47) (19.21) (17.54) (14.14)

chi2 for
models 4895.0*** 2832.6*** 1591.3*** 2861.4*** 1743.0*** 2216.9*** 680.3*** 892.3*** 734.4***
significance

observations 85281 38172 47109 53940 31341 43916 21686 13089 9141

Note: t statistics appear in parentheses. ***, **, * indicate statistical significance at the 0,1%; 1% and 5% level respectively. Regressions
include country and industry fixed effects as well as a constant term which are not reported.
Source: Community Innovation Survey (CIS)

145
Table A9. Results of the R&D regressions with CIS2008
Medium-
Small Large All manuf. Medium-
All firms EU-15 EU-12 Low-tech high and
firms firms firms low tech
high tech

R&D expenditures equation

extramural
0.619*** 0.556*** 0.729*** 0.510*** 0.633*** 0.702*** 0.745*** 0.765*** 0.542***
R&D

(13.70) (11.64) (7.81) (8.67) (9.41) (11.99) (7.31) (8.31) (4.80)

vertical
0.559*** 0.437*** 0.769*** 0.540*** 0.567*** 0.492*** 0.807*** 0.302** 0.424**
collaboration

(9.98) (7.26) (6.82) (7.34) (6.87) (6.78) (6.24) (2.64) (3.11)

horizontal
0.121 0.170* -0.000436 0.162 0.0841 0.129 -0.117 0.364* 0.0484
collaboration

(1.74) (2.10) (-0.00) (1.66) (0.86) (1.35) (-0.67) (2.46) (0.27)

intra-group
0.514*** 0.686*** 0.23 0.267* 0.454*** 0.588*** 0.553** 0.571*** 0.600***
collaboration

(7.30) (8.55) (1.79) (2.30) (4.97) (6.42) (3.16) (3.97) (3.66)

local public
0.219*** 0.228*** 0.218 0.250** 0.209* 0.157 0.0257 0.158 0.292*
funding

(3.44) (4.01) (0.78) (3.26) (2.05) (1.93) (0.17) (1.24) (1.97)

national
public 1.032*** 1.004*** 1.098*** 0.953*** 1.044*** 1.011*** 1.080*** 1.037*** 0.890***
funding

(19.08) (18.19) (9.05) (13.22) (13.19) (14.75) (8.44) (9.70) (7.17)

EU funding 0.716*** 0.960*** 0.440** 0.740*** 0.693*** 0.666*** 0.766*** 0.533** 0.748***

(8.77) (9.61) (3.12) (6.25) (6.14) (6.20) (4.08) (3.04) (3.76)

firm size
0.957*** 0.819*** 1.151*** 0.907*** 0.722*** 1.075*** 0.923***
class

(18.40) (15.24) (9.96) (13.22) (6.12) (10.13) (6.76)

R&D selection equation

log 2006
0.0380*** 0.0253*** 0.0880*** 0.00792*** 0.108*** 0.0532*** 0.0649*** 0.0436*** 0.0533***
sales

(23.99) (14.27) (23.92) (4.37) (24.77) (22.83) (15.69) (11.99) (11.63)

member of a
0.450*** 0.431*** 0.451*** 0.390*** 0.289*** 0.529*** 0.543*** 0.562*** 0.462***
group

(40.14) (29.65) (24.88) (22.19) (17.63) (33.32) (20.80) (21.69) (14.58)

chi2 for
models 4794.1*** 3828.8*** 1004.0*** 2795.5*** 1862.0*** 2322.88*** 829.8*** 958.0*** 596.3***
significance

observations 98345 48831 49514 60845 37500 47306 23667 15152 8487

Note: t statistics appear in parentheses. ***, **, *indicate statistical significance at the 0.1%; 1% and 5% level respectively. Regressions
include country and industry fixed effects as well as a constant term which are not reported.
Source: Community Innovation Survey (CIS)

146

EU PRODUCTION AND TRADE BASED ON KEY ENABLING
BACKGROUND
Previous chapters have discussed the specialisation,
complexity and sophistication of economies basing
their output on key enabling technologies (KETs).
This chapter takes an in-depth look at the
specialisation, strengths and weaknesses of the EU in
the global production and trade in products based on
KETs.
Two years ago, the High-Level Group on Key
Enabling Technologies published its final report
which estimated that the global market potential for
products based on KETs would grow from USD
832 bn around 2008 to USD 1,282 bn around 2015
(HLG KETs 2011).
It was followed by the European Commission
Communication A European strategy for Key
Enabling Technologies A bridge to growth and
jobs (European Commission 2012 a) which outlined
a strategy to boost the industrial production of KETs-
based products and enable maximum exploitation of
the EUs potential in competitive markets.
In addition, in its Communication A stronger
European Industry for Growth and Economic
Recovery (European Commission 2012 b), the
Commission identified six priority action lines, one of
which was the creation of markets for KETs. The
European Commission expressed its intention to
implement the European Strategy for KETs, ensuring
better co-ordination of EU and Member State
technology policies; funding of essential
demonstration and pilot lines and cross-cutting KET
projects; and the timely development of the internal
market for KETs-based products (Calleja 2013).
Moreover, the industrial deployment of KETs will be
considered in future European Innovation
Partnerships, while a knowledge and innovation
community on added-value manufacturing has been
proposed as a forum for integration and promotion of
skills and competences (European Commission
2013 c).
OBJECTIVES
The overall objective of this chapter is to analyse the
current position of the EU in the global production of
KETs-based products in order to assess upcoming
challenges for the competitiveness of the EU. The
chapter aims to:
147
Chapter 5.
TECHNOLOGIES
Provide a narrative overview of most recent
technological and industry developments in each
KET since 2009;
Update estimations on future market potentials in
each KET, building on the analyses of recent
trends in market shares in the production of
KET-related technologies;
Assess the EU position in the value chain by
studying two promising KETs-based products;
Analyse the EU position in international trade for
certain subfields of KETs-based products,
including changes in the competitiveness of the
EU over time;
Determine the EU position in value chains (in
terms of technology content) within certain
subfields of KETs-based products based on unit
value analysis of exports and imports;
Analyse the specialisation of a selection of EU
Member States in production and trade of KETs-
based products by combining production and trade
statistics.
This chapter applies the following definition of
KETs-based products (European Commission
2012 a). A KETs-based product is: (a) an enabling
product for the development of goods and services
enhancing their overall commercial and social value;
(b) induced by constituent parts that are based on
nanotechnology, micro-/nanoelectronics, industrial
biotechnology, advanced materials and/or photonics;
and, but not limited to (c) produced by advanced
manufacturing technologies.
STRUCTURE OF THE CHAPTER
This chapter is structured as follows: Section 5.1
presents an update of market share calculations and
market potential estimates. 5.2 analyses the position
of the EU in international trade in KETs-based
products. In 5.3 to 5.5, the value chain of two KETs-
based products is analysed, namely lipase enzymes
and the accelerometer. Section 5.6 summarises the
main conclusions and potential policy implications.
5.1. TECHNOLOGY POSITIONS AND MARKET
POTENTIAL
5.1.1. Introduction
Key Enabling Technologies (KETs) are defined as
knowledge-intensive technologies associated with
high R&D intensity, rapid innovation cycles, high
capital expenditure and highly skilled employment.
They are multidisciplinary, cutting across many
technology areas with a trend towards convergence
and integration.
The following technologies are identified as KETs:
micro- and nanoelectronics, nanotechnology,
photonics, advanced materials, industrial
biotechnology and advanced manufacturing
technologies for other KETs (HLG KETs 2011).
The objective of this section is to provide an
overview of the competitive position of the EU in the
generation of technology and to estimate the future
market potential for KETs-based products and
applications. As such, it provides an update of the
analysis undertaken in the background study to the
2010 European Competitiveness Report (European
Commission 2010). The calculation of technology
market shares is based on the number of international
patent applications. KETs-relevant patent activities
are identified through a list of IPC codes developed
for the 2010 report and recently updated in the
Feasibility study for a KETs Observatory
commissioned by DG Enterprise and Industry (Van
de Velde et al. 2013). In order to estimate the market
potential of each KET, an analysis of existing studies,
reports and reviews has been conducted. For each
KET, several market segments have been selected,
depending on KETs-based applications.
5.1.2. Approach
An important measure of a countrys competitive
position in KETs is its ability to produce new,
commercially relevant technological knowledge. One
way to measure this ability is to look at patent data.
Patent data have certain advantages when it comes to
measuring technological performance. Patents
represent new technological knowledge that has a
particular potential for economic application. Each
patent is linked to technological areas through an
internationally standardised system (International
Patent Classification (IPC)) which enables patents to
be linked to KETs. Since patents are essential for
the production and protection of new technologies
and innovative products and processes, they are a
commercial good which serves as an input to
production and can be traded on technology markets
(through licensing or by selling and purchasing patent
rights). In contrast to many other goods, most patents
are produced and used in-house while only a small
part is actually traded between firms (see
Gambardella et al. 2007; Arora et al. 2002; Serrano
2005; Lamoreaux and Sokoloff 1999).
When using patent applications to assess the
competitive strength and weakness of an economy,
some limitations need to be pointed out. First, not all
new technological knowledge needed for innovations
148
is represented by patents, while a number of patents
will never be used for innovations. Secondly, the
economic value represented by one patent can vary
substantially. Thirdly, not all patents seek legal
protection of new technological knowledge but some
are used to block competitors from patenting
activities or to keep strategic information away from
competitors. For these reasons, patents represent only
a fraction of the technology market.
As with any other market, one can analyse the
technology market performance of individual actors
as well as of countries. Here, for each country a
market share of the technology market for each
KET is calculated based on the number of
international patent applications. International patent
applications are patents applied for at the European
Patent Office (EPO) or through the Patent
Cooperation Treaty (PCT) procedure at the World
Intellectual Property Organization. Using
international patent applications reduces the risk of an
overly strong home-country bias and excludes patents
of low (expected) commercial value since applying at
the EPO or via the PCT is comparatively costly.
Technology market shares by KET are calculated
using a conversion table that links IPC codes to KETs
(see Van de Velde et al. 2013). Patent applications are
assigned to countries using the country of the
applicant and by applying fractional counting in the
event that a patent application is submitted by
organisations from different countries. Patents are
assigned to four regions: Europe (all EU Member
States plus Albania, Andorra, Bosnia-Herzegovina,
Former Yugoslav Republic of Macedonia, Iceland,
Liechtenstein, Monaco, Montenegro, Norway, San
Marino, Serbia, Switzerland); North America (US,
Canada, Mexico); East Asia (Japan, China including
Hong Kong, South Korea, Singapore, Taiwan); and
the rest of the world (RoW). The April 2013 edition
of the Patstat database published by EPO is used.
5.1.3. Industrial biotechnology
5.1.3.1. Technology market share
International patent applications in the field of
industrial biotechnology have been decreasing over
the past ten years. Globally, the number of patents fell
by 33 % between 2000 and 2010 (Figure 5.1). Europe
and North America report even greater drops ( 46 %).
East Asia and RoW increased the number of
international patent applications in industrial
biotechnology by 28 % and 14 % respectively. As a
consequence, the market shares of Europe and North
America are declining. Nevertheless, North America
remains the region with the highest market share in
2010 (39 %). Europe lost its second position in 2010
even though its market share in that year (27 %) was
above the low level reported for the mid-2000s (23 %
in 2006). East Asia gained market shares and
contributed 28 % to global patent applications in
industrial biotechnology in 2010. Rest-of-the-world
countries showed increasing market shares up to 2008
but no further growth afterwards, contributing 5 % to
total patenting in industrial biotechnology in 2010.
In Europe, Germany gradually lost market share,
declining from 44 % (2000) to 27 % (2010). France
gained market shares and by 2008 had replaced the
UK as the second largest European patent producer in
industrial biotechnology. The Netherlands showed
high market shares in the mid-2000s (ranking second
in 2005 with a European market share of 15 %) but
clearly lost ground in recent years. Switzerland and
Denmark hold position five and six in European
patenting in industrial biotechnology.
5.1.3.2. Market potential
Industrial biotechnology is used in the production of
chemicals and derived biomaterials. The use of
biotechnology for chemical production has increased
over the past decade and is likely to continue
increasing, driven by rising energy costs, new
chemicals legislation and increasingly stringent
environmental regulations (OECD 2009).
According to Festel Capital, the sales of products
made by biotechnological processes in 2007 was
around EUR 48 bn, or 3.5 % of total chemical sales,
while by 2017 predicted sales of products made by
biotechnological processes will be around EUR
340 bn, or 15.4 % of total chemical sales in 2017.
Based on Festel Capital research, the most important
sub-segments in 2017 are expected to be active
pharma ingredients and polymers and fibres (Festel
2010). Other sources start from a market share of 9
13 % in 2010 and predict further growth to 22 28 %
by 2025. Major growth is expected to take place in
polymers and bulk chemicals (Kircher 2012).
The global market for industrial enzymes is forecast
to reach USD 3.74 bn by 2015. Important factors
driving the market include new enzyme technologies
with a view to enhanced cost efficiencies and
productivity, and growing interest in substituting
petroleum-based products. BCC projects the
industrial enzymes market to grow to USD 6 bn by
2016 (BCC Research 2011 a). Major growth is
expected in the segments of food and beverage
enzymes and technical enzymes. Two other segments
with high growth potential are carbohydrases and
lipases (see also 5.4).
5.1.4. Photonics
5.1.4.1. Technology market share
Over the past ten years, East Asia has gained
significantly in technology market shares in the field
of photonics (Figure 5.2). Since 2003, East Asian
organisations have become the largest group of
149
applicants for photonics patents and have been able to
strengthen their position continuously, increasing
their market share from 27 % in 2000 to 50 % in 2010.
North American applicants lost the leading position
which they held in the early 2000s. Their market
share fell from 40 % (2000) to 19 % in 2010. Europe
did significantly better: its market share increased
until 2008, when it reached 32 %. In 2009 and 2010,
Europes contribution to photonics patenting fell back
to 29 %. Countries from outside the three main
regions slightly lost market shares.
Changes in market shares in photonics took place
against the background of expanding overall
patenting. The total number of international patent
applications grew by 25 % between 2000 and 2010,
almost four times the growth rate for all KET patent
applications and equal to the growth rate of patenting
across all fields of technology.
Germany further strengthened its position as the main
producer of new technological knowledge in
photonics within Europe over the past decade. Its
share of total European patent applications was 43 %
in 2010, compared to 33 % in 2001 2002. Among the
other five main European applicant countries in 2010
France, Netherlands, UK, Austria and Italy
Netherlands and the UK lost market shares while
France and Italy maintained their positions within
Europe. Austria recently increased patenting in
photonics and overtook Swiss patents applicants.
5.1.4.2. Market potential
The photonics industry is expected to grow
significantly in coming years. The global market for
photonic components and systems forecast to be
worth EUR 480 bn by 2015, suggesting an annual
growth rate of 8 % (HLG KETs 2011).
Solar photovoltaic (PV) is the third most important
renewable energy in terms of globally installed
capacity. Its growth rate reached almost 70 % in 2011.
In terms of cumulative installed capacity, Europe
leads the way worldwide with more than 51 GW
installed as of 2011 (75 % of world capacity).
Internationally, significant market growth is expected
until 2017, reflecting the large untapped potential of
many countries (EPIA 2013).
Figure 5.1. Market shares in international patent applications in industrial biotechnology, 20002010 (percent)
Industrial Biotechnology - global 55 Industrial Biotechnology - Europe

55 Europe North America East Asia RoW 50 DE FR NL GB CH DK

market share in European patent applications, %

50 45
market share in all patent applications, %

45
40
40
35
35
30
30
25
25
20
20
15
15

10 10

5 5

0 0
'00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10

Source: EPO: Patstat, ZEW calculations

By 2020, light emitting diodes (LEDs) are expected
to account for around 95 % of the market for light
bulbs, currently estimated at EUR 11 bn per year (J.P.
Morgan Cazenove 2012). The expected growth in
market demand for LEDs will be driven by product
substitution. Other application areas of LEDs are:
mobile applications including mobile phone
notebooks and tablets; TV and monitor backlights;
sign and automotive lighting. Japan accounts for the
greatest portion of overall LED component revenues
(30 %) followed by South Korea (26 %), Taiwan and
Southeast Asia (19 %).
The optical communication industry is experiencing a
recovery from the economic downturn. In 2010 and
2011, the sales of data communication systems
started to pick up again. The global market for lasers
for communications (data and telecoms) was
estimated to be worth USD 1.95 bn in 2010 and USD
2.22 bn in 2011 (+ 14 %) (Overton et al. 2011). While
Europe is experiencing a decline in demand, the
construction of optical communication is at a peak in
China.
5.1.5. Micro-/nanoelectronics
5.1.5.1.
East Asia has since 2002 been the largest producer of
international micro- and nanoelectronics patent
applications (Figure 5.3). Its market share is gradually
increasing over time. In 2010, 56 % of global patent
applications in this KET originated in East Asia.
North America and Europe are both losing market
shares. In 2010, North America reported a market
share of 23 %, while the figure for Europe was 20 %.
Countries from the rest of the world are of little
importance in the technology market for micro- and
nanoelectronics: their market share is 1% to 2 %.
Dynamics in micro- and nanoelectronics patenting are
high. Globally, patent applications grew by 35 % from
2000 to 2010. The number of European applications
in 2010 was 2 % higher than in 2000, while applicants
from North America reported a 17 % lower figure in
2010 than in 2000. The highest growth is in East
Asia, where patent applications increased by 116 %
over the same ten-year period.
In Europe, Germany is clearly the largest patent
producer in micro- and nanoelectronics and
maintained a European market share of 42 45 % from
2001 to 2010. France increased its European market
share from 10 % (2000) to 17 % (2010), overtaking
the Netherlands. The Dutch market share within
Europe declined from 19 % in 2003 to 8 % in 2010.
The UK is the fourth largest producer of micro-
/nanoelectronics patent applications in Europe,
followed by Switzerland and Italy.
5.1.5.2. Market potential
The global market for the semiconductor industry has
increased significantly from USD 25 bn in 1985 to
USD 299.5 bn in 2011 (SIA 2012). This growth is
driven by the increasing need for microelectronic
devices and smart sensors in intelligent products,
such as smart phones, tablets, car driver assistance
Technology market share systems, smart grids, networked sensors, and other
products. Smart sensors can also be used to detect and
make risk assessments of disasters. That sort of risk
management reduces the vulnerability of Member
States, sectors and individual firms, thereby
increasing competitiveness and sustainable growth.

150
Figure 5.2. Market shares in international patent applications in photonics, 20002010 (percent)

Photonics - global 55 Photonics - Europe

55
Europe North America East Asia RoW 50 DE FR NL GB AT IT

market share in European patent applications, %

market share in global patent applications, %

50
45
45
40
40
35
35
30
30
25
25
20
20
15
15

10 10

5 5

0 0
'00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10

Source: EPO: Patstat, ZEW calculations

Figure 5.3. Market shares in international patent applications in micro and nanoelectronics, 20002010 (percent)
Micro- and nanoelectronics - global Micro- and nanoelectronics - Europe
55 Europe North America East Asia RoW 55
DE FR NL GB CH IT

50 50
market share in European patent applications, %
market share in global patent applications, %

45 45

40 40

35 35

30 30

25 25

20 20

15 15

10 10

5 5

0 0
'00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10

Source: EPO: Patstat, ZEW calculations

From a total investment of EUR 28 bn in segment of the market and global economic
microelectronics in 2007, only 10 % was in the EU, uncertainty (Clarke 2013). The strongest growth is
compared to 48 % in Asia. Europes semiconductor expected for microprocessor units, especially in the
market share has declined from 21 % to 16 % since area of tablet computers and smartphones.
2000 (Silicon Europe 2012). After the global
economic crisis, the semiconductor market recovered The total flash memory market grew by 2 % to USD
quickly and global sales reached a record high in 30.4 bn by end-2012, overtaking the DRAM 139
2010. While billings fell by 11 % from their peak in market for the first time, as the latter declined from
2007 to 2009, sales subsequently recovered by 33 % USD 31.2 bn to USD 28 bn. This is because DRAM is
from 2009 to 2010, an unprecedented growth rate used mostly in PCs while flash memory is used in
which more than compensated for previous losses smartphones, media tablets, and other personal media
(Ballhaus et al. 2011). PWC estimates that the devices. IC Insights forecasts NAND 140 flash
semiconductor market will grow by 7.4 % per year on memory sales to increase by 14 % annually from 2012
average from 2010 to 2015. to 2017, growing to USD 53.2 bn by 2017, while the
DRAM market is forecast to grow by 9 % over the
According to IC Insights, worldwide processor sales same period.
are expected to regain strength in 2013 and grow 12 %
to USD 65.3 bn, after a more modest increase in 2012
to USD 58.2 bn (+ 5 %). The slow growth in 2012 is 139
Dynamic random-access memory.
attributed to weaknesses in the personal computer 140
The other main type of flash memory is NOR.

151
5.1.6. Advanced materials
5.1.6.1. Technology market share
East Asia is constantly increasing its market share in
international patenting in the field of advanced
materials (Figure 5.4). In 2010, 48 % of all advanced
materials patent applications originated in East Asia,
compared to 28 % from Europe and 21 % from North
America.
North Americas share of global patent applications is
declining much faster than the European share.
Changes in market shares should be seen against the
backdrop of low patent dynamics in advanced
materials. Global patent applications fell by 4 %
between 2000 and 2010. While international patent
applications in advanced materials are going down in
Europe and North America, applicants from East
Asian and the rest of the world are filing more
applications each year.
Within Europe, the market shares of countries with
the highest numbers of advanced materials patents
have remained stable over time. Germany still
accounts for more than 40 % of European patent
applications, followed by France (16 % in 2010),
Italy, Switzerland, the UK and Belgium. The
Netherlands held third place in advanced materials
patenting in Europe until 2008 but its patent activities
have since decreased considerably.
5.1.6.2. Market potential
Advanced materials tend to outperform conventional
materials with their superior properties such as
toughness, hardness, durability and elasticity. The
scope of advanced materials research is very broad.
While some advanced materials are already well-
known, like polymers, metal alloys, ceramics,
semiconductors, composites and biomaterials, other
advanced materials like carbon nanomaterials,
activated carbon, titanium, are becoming increasingly
important.
Smart materials are a class of materials that respond
dynamically to electrical, thermal, chemical,
magnetic, or other stimuli from the environment.
These materials are incorporated in a growing range
of products, enabling these products to alter their
characteristics or otherwise respond to external
stimuli. The market for these materials was estimated
to be worth USD 19.6 bn in 2010 and was expected to
approach USD 22 bn in 2011 and exceed USD 40 bn
by 2016, a compound annual growth rate (CAGR) of
12.8 % from 2011 to 2016 (BCC Research 2011 b).
152
Lightweight materials are increasingly being used in
the transportation industry as weight reduction is one
of the most important ways of reducing fuel
consumption. In 2010, the total global consumption
of lightweight materials used in transportation
equipment was worth USD 95.5 bn. By 2015 this
market is expected to reach USD 125.3 bn, with a
compound annual growth rate (CAGR) of 5.6 %
between 2010 and 2015.
Value-added materials (VAMs) are a group of
advanced materials with strategic importance for
economic growth, industrial competitiveness and
societal challenges. Their market potential is
estimated to reach EUR 1,000 bn by 2050. In the
environmental market segment, VAM growth will be
driven by energy-efficient and carbon-capture
technologies. VAMs in the ICT sector are expected to
grow substantially in the coming years, with an
average compound annual growth rate of 5 %.
5.1.7. Nanotechnology
5.1.7.1. Technology market share
Trends in technology market shares in the field of
nanotechnology significantly diverge from the
general trends in KETs patenting. With a share of
39 % of all applications in 2010, North America is
still the most important origin of nanotechnology
patent applications. While North Americas share of
all nanotechnology applications was falling until
2007 (when it reached 35 %), the downward trend
changed in 2008.
Europe and East Asia report similar market shares
over the entire period. In most years, the East Asian
share of applications exceeded the European share but
in recent years Europe has taken a slightly higher
share (28 % in 2010, versus 27 % for East Asia). The
total number of nanotechnology patent applications
grew by 31 % between 2000 and 2010, with all four
regions reporting growing nanotechnology patenting.
Within Europe, Germany has lost market shares over
the past decade, from 41 % in 2000 to 23 % in 2010.
At the same time, France substantially increased its
nanotechnology patenting and gained market shares,
catching up with Germany in 2010. The market share
of the Netherlands dropped from 14 % in 2004 to 6 %
in 2010, while the UK was able to maintain its share
of total European patent applications in
nanotechnology at around 10 %. Switzerland filed
about 5 % of European nanotechnology patent
applications over the entire period, while Italy has
recently increased its share.
Figure 5.4. Market shares in international patent applications in advanced materials, 20002010 (percent)

55 Advanced Materials - global 55 Advanced Materials - Europe

50

market share in European patent applications, %

50
market share in global patent applications, %

Europe North America East Asia RoW DE FR BE GB CH IT

45 45

40 40

35 35

30 30

25 25

20 20

15 15

10 10

5 5

0 0
'00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10

Source: EPO: Patstat, ZEW calculations

the quantum dots technology market, followed by

5.1.7.2.
Nanotechnology has many applications in a broad
range of industries. The global market for
nanotechnology was valued at USD 20.1 bn in 2011
and USD 20.7 bn in 2012 (BCC Research 2012).
Total sales are expected to reach USD 48.9 bn in 2017
after increasing at a five-year compound annual
growth rate of 18.7 %. The US is the most prominent
market and in 2011 accounted for an estimated share
of around 35 % of the global nanotechnology market
slightly less than its share of patent applications.
Whilst it is expected to remain a major player,
emerging economies such as China and South Korea
as well as India and Brazil have started to catch up.
The global market for products based on the
revolutionary new nanomaterial graphene is projected
to reach USD 122.9 million in 2017 and USD 986.7
million in 2022, growing at a five-year compound
annual growth rate of 51.7 %. The segment made up
of capacitors is projected to be the largest segment in
2022. Capacitors are expected to increase from USD
31 million in 2017 to USD 410 million in 2022, a
CAGR of 67.6 %. Others sources indicate a more
conservative estimate of USD 100 million in 2018
and an annual growth rate of 40 %, making the
capacitors segment worth USD 216 million by 2020.
The global market for quantum dots, which in 2010
was generated revenues of USD 67 million, is
projected to grow over the next five years at a
compound annual growth rate of 58.3 %, reaching
almost USD 670 million by 2015 a tenfold increase.
MarketsandMarkets estimate the total market for
quantum dots to be worth USD 7.5 bn by 2022, the
result of a compound annual growth rate of 55.2 %
from 2012 to 2022. The US has a leading position in
Market potential Europe and Asia-Pacific (MarketsandMarkets 2012).
5.1.8. Advanced manufacturing technologies
5.1.8.1. Technology market share
Trends in market shares for advanced manufacturing
technologies for other KETs are quite similar to those
for micro-/nanoelectronics and advanced materials,
since many patents classified as advanced
manufacturing technologies for other KETs relate to
the former two KETs and a significant overlap exists.
East Asia is producing the highest number of patents
in the field of advanced manufacturing technologies
for other KETs (46 % in 2010), while Europe and
North America have about 25 % each of global patent
applications. North Americas contribution to the
global patent output has fallen sharper (from 40 % in
2000) than Europes share (31 % in 2000). Rest-of-
the-world countries increased their share marginally
between 2000 and 2010, contributing 3 % to global
patent applications in 2010.
Patent dynamics in this KET are low. The total
number of international patent applications in 2010
was 9 % below the 2000 figure. Declining patent
output in Europe ( 25 %) and North America ( 41 %)
are partly outweighed by significant increases in East
Asia (+ 57 %) and RoW (+ 11 %).
In Europe, Germany has lost market shares but is still
the largest patent producer in this KET with a
European market share of 38 % in 2010. France
follows second with 17 % to European patent output
in in 2010. The Netherlands has fallen to rank 5 in
2010, overtaken by the UK and Switzerland. Sweden
was the sixth largest patent producer in Europe in
2010, ousting Italy to rank 7.

153
5.1.8.2. Market potential
Manufacturing is an essential step to bring
technological innovations to the market. The global
manufacturing economy is estimated to be worth
GBP 6.5 trillion (TSB 2012). In the 2013 Global
Manufacturing Competitiveness Index, China was
found to be the most competitive manufacturing
nation, followed by Germany, US, India and South
Korea. Five years from now, the report predicts China
to maintain the first ranking, followed by India,
Brazil, Germany and US (Deloitte 2013).
Additive manufacturing is a layer-by-layer technique
of producing three-dimensional objects directly from
a digital model. With markets such as prototyping,
tooling, direct part manufacturing, and maintenance
and repair, the industry has grown significantly to
USD 1.3 bn of materials, equipment, and services in
2010. The additive manufacturing market, including
consumer products, business machines, medical, and
aerospace industries, is expected to grow at a
compound annual growth rate (CAGR) of 13.5 %
from 2012 to 2017.
In 2011, BCC Research estimated the global market
for robots and robot-related products to grow to
nearly USD 22 bn in 2011 and USD 30 bn by 2016, a
compound annual growth rate (CAGR) of 6.7 %. In a
more recent report (BCC Research 2013), BCC
forecast a slightly lower compound annual growth
rate (CAGR) of 5.9 % between 2013 and 2018. The
Asian market is expected to see the fastest growth in
the coming years, while growth in the European
Union is anticipated to be concentrated in the latter
part of the forecast period, when robotic development
initiatives now being undertaken on an EU-wide basis
will result in commercialised products.
5.1.9. North American decline, East Asian rise
The preceding analysis of shares of patent
applications for KETs reveals a steady strengthening
of East Asia as the main producer of new
technological knowledge in KETs. Over the past ten
years, East Asian organisations have increased their
share in total patent activity in each of the six KETs.
In four KETs photonics, advanced materials, micro-
/nanoelectronics, advanced manufacturing
technologies for other KETs East Asian applicants
were the most important patent producers by 2010. At
the beginning of the 2000s, North America held the
leading position in all six KETs: nowadays industrial
biotechnology and nanotechnology are the only two
areas that still show North America as the region with
the largest share of patent applications. While North
America has lost market shares in all six KETs,
Europe has performed relatively better. In photonics
and nanotechnology, Europe has remained stable
during the past decade, while in the other four KETs
losses were less severe compared to North America.
154
The decline in North America and general stability in
Europe occurred despite productivity gains in North
American manufacturing which have tended to be
greater than in Europe.
These trends in KET patenting are very similar to the
overall trend in international patenting: an increasing
East Asian share and a declining contribution by
North America, while Europe reports moderate losses
in market shares. The main difference with respect to
KETs is the speed with which East Asia captures
market shares, giving this region a leading position
globally. By contrast, the shift from West to East in
general international patenting is taking place more
slowly, with Europe still holding the largest share in
2010.
In Europe, Germany and France were the main
sources of patent applications in 2010 in each of the
six KETs. While Germany maintained its dominant
position during the past ten years, France increased its
share of total European patent applications in all six
KETs. The UK and Netherlands both show
decreasing market shares.
A more disaggregated analysis at the level of
subfields within each KET reveals that Europe is the
leading KET patent applicant in some subfields and
has been able to gain market shares. In photonics,
European strengths are in the fields of measurement
and electro-optics as well as lasers. In
nanotechnology, Europe is the leading source of
patent applications in nano-analytics and has
increased its market share in nano-materials. In
micro- and nanoelectronics, Europe has been able to
maintain its market share in the field of devices and
shows an increasing market share in the small area of
testing and amplifiers. In advanced manufacturing
technologies for other KETs, Europe has a very high
market share in the subfield of instruments and has
been able to maintain its share in the global
technology output of advanced manufacturing
technologies for biotechnology and materials
production.
The analysis of the market potential of KETs reveals
that substantial market growth is expected in all six
KETs over the coming years. Depending on the KET,
growth potentials of 10 20 % per year can be
expected. For particular submarkets, the growth
potential is even larger. The position of Europe with
respect to market size differs for the various KETs,
but in general the increasing importance of East Asia
and the higher pace of market share gains can be seen
here as well.
5.2. THE POSITION OF EUROPE IN THE
PRODUCTION AND TRADE OF KETS-RELATED
PRODUCTS
5.2.1. Introduction
Analysing the position of countries and regions
within value chains of a certain production process
typically requires information on input and output
links between the countries or regions. Input-output
tables, however, offer such information only at a
highly aggregated level of industries, not for
individual products which can be linked to KETs. For
that reason, this section uses an alternative approach.
In order to identify Europes position in global value
chains within each KET in relation to North America
and East Asia, characteristics of production and trade
and their relation to technology inputs are examined.
The following metrics are complementary and will be
used jointly:
1. The technology content of manufactured goods,
i.e. whether products are more technologically
advanced;
2. The type of competition of Europes exports in
KET-related products, distinguishing between
quality and price competition;
3. The links between the creation of new
technological knowledge (measured by patent
applications) and the technology content of
manufactured goods.
By combining these three approaches, a
comprehensive picture of Europes competitive
advantage vis--vis its main competitors in each KET
will emerge. The analysis is based on data for
individual products related to one of the six KETs, in
the sense that the products represent certain
technological features which are directly linked to a
KET (a certain new material, a photonics element, a
semiconductor, a biochemical entity or a machine
tool) but which do not use KETs as an input for more
complex goods (such as batteries, measuring
instruments, medical devices, information and
communication devices). The notion of value chain
as used in this chapter therefore refers entirely to the
division of labour within the production of KET
products. In order to identify products linked to
KETs, the results of a recent feasibility study on
monitoring KETs are used (Van de Velde et al. 2013).
155
In that study, KET products were defined at an 8-digit
level of the Prodcom product classification system.
For the purposes of this report, a narrow version of
the definition is used in order to avoid analysis of
products that are only partially linked to a certain
KET.
5.2.2. Technology content of products related to
key enabling technologies
The concept of technology content assumes that
similar products can be produced by using different
qualities and quantities of technology. Technology
may refer to the sophistication of production
methods, the variety of different technologies used in
the production, or how technologically advanced
inputs are. Products with higher technology content
are supposed to be positioned further along the value
chain. As high technology content products should be
superior to products with lower technology content,
they should also reflect a higher unit price. Therefore
a common trade indicator will be used to measure
technology content: the unit value of exports. Based
on the assumption that a countrys exports of a certain
product represent that countrys total production of
the product, export unit values give the average value
of a product manufactured in a country. The
assumption is somewhat unrealistic, since many
studies have shown that exports tend to contain more
innovative products than the average since it is the
more innovative firms that engage in exports (see
Wakelin 1998; Bleaney and Wakelin 2002; Beise and
Rammer 2006; Wagner 1996; Ebling and Janz 1999;
Roper and Love 2002; Lefebvre et al. 1998). Here
though, the possible bias of exports towards
innovative products can be seen as an advantage
because it means the analysis will focus on the more
innovative products within each KET.
A country or regions export unit value of a certain
product is compared with the export unit value of the
same product in global trade. A value greater than one
indicates that the country (region) exports (and
therefore manufactures) products of a higher value
per unit, hence products with a higher technology
content. Comparing export unit values over time
provides information about the dynamics in
technology content, in other words whether a country
(region) moves away from the average unit value or
converges towards it. Combining both dimensions
the level and dynamics of unit values produces a
matrix with four quadrants.
Figure 5.5. Measuring technology content of manufactured products
p=1: high and increasing p=3: low but increasing
technology content technology content
UVXijt > UVXwjt UVXijt < UVXwjt
UVXijt > UVXwjt UVXijt > UVXwjt
p=2: high but decreasing p=4: low and decreasing
technology content technology content
UVXijt > UVXwjt UVXijt < UVXwjt
UVXijt < UVXwjt UVXijt < UVXwjt
TC(p)ikt = j Xpij(k)t / ip Xpij(k)t
For each KET-related product, a country (region) can be positioned in the following way:
1. High and increasing technology content. 3.
2. High but decreasing technology content. 4.
Source: NIW/ZEW
The technology content (TC) of a country (region) i
in a certain KET k is examined by determining the
position p in the quadrants shown in Figure 5.5 (p
{1,2,3,4}) of each individual product j belonging to
KET area k, weighted by the products share in total
exports X of products related to KET area k of
country (region) i.
The analysis is conducted for three regions: EU-28
(EU Member States), North America (US and
Canada) and East Asia (Japan, South Korea and
China). The total exports of the three regions
constitute total trade w. Furthermore, separate
analyses for 12 Member States (Germany, France,
United Kingdom, Italy, Belgium, Netherlands,
Austria, Sweden, Denmark, Poland, Czech Republic
and Hungary) are carried out. The analysis is
undertaken for individual KET-related products
defined as 6-digit classes of the HS (harmonised
system) product classification used in trade statistics.
The 6-digit HS classes were identified using a
conversion table from 8-digit Prodcom codes. Data
on exports (in USD) and quantities (kg) were taken
from the UN Comtrade database. The analysis covers
the period from 2002 to 2011. Data for 2007 to 2011
rely on the HS 2007 classification while data for 2002
to 2006 are based on HS 2002. A conversion table
was used to link the two classifications. To avoid
picking up unit value fluctuations between single
years, the analysis focuses on the development
between two sub-periods, 2002 2006 and 2007
2011. In order to classify products by their
technology content, changes in unit values between
the average values for 2002 2006 and 2007 2011
are calculated.
The EU-28 reports a high and increasing
technological content of its exports in four KETs. The
strongest performance is found for industrial
biotechnology. Here about 90 % of the exports in the
years 2007 to 2011 were generated by products with a
higher unit value than in the main competitor regions
156
UV ... unit value ($/kg)
X ... exports
... change over time
i ... country/region (i w)
w ... world (total trade)
j ... KET-related product
k ... KET
t ... time
TC ... technology content
p ... position in quadrant
Low but increasing technology content.
Low and decreasing technology content.
and for which unit values increased more rapidly over
time than in the other regions (Figure 5.6). In
advanced manufacturing technologies (AMT) for
other KETs, 74 % of EU-28 exports were in products
with high and increasing technology content. For
photonics exports, the corresponding share was 69 %
and for advanced materials it was 57 %.
Low technology content occurs for nanotechnology
and micro- and nanoelectronics. In nanotechnology,
28 % of the EU-28 export volume 2007 2011 was
generated by products with high and increasing unit
values, while 37 % of exports were in products with
low and decreasing technology content and another
30 % with low but decreasing unit values. For micro-
and nanoelectronics, because of limited data
availability only the 2002 2006 period can be
analysed. In this earlier period, the EU-28 exported
products related to micro- and nanoelectronics with a
lower unit value compared to the same products
exported by the main competitor regions (East Asia in
particular). However, for almost all of these products,
unit values increased more strongly in the first half of
the 2000s (from 2002 2003 to 2005 2006) than in
the competitor regions.
Whilst North America shows a similar pattern of
technology content of exports of KETs-related
products as the EU-28, it performs significantly better
(in terms of having a higher technology content of
exports) in photonics and nanotechnology but less
well in industrial biotechnology and advanced
materials. In photonics, North America is the leading
technology region with broad and constantly
increasing technological content. In nanotechnology,
75 % of the 2007 2011 exports were based on
products with high technology content, and for most
of these products, unit values increased more rapidly
than in the competitor regions.
Figure 5.6. Technology content of KET-related products by KET and triadic region, 2007 2011 averages

high and increasing technology content high but decreasing technology content
low but increasing technology content low and decreasing technology content
EU-28
Industrial Biotechnology
Nanotechnology
Photonics
Micro- and Nanoelectronics*
Advanced Materials
AMT for other KETs

0 10 20 30 40 50 60 70 80 90 100

North America
Industrial Biotechnology
Nanotechnology
Photonics
Micro- and Nanoelectronics*
Advanced Materials
AMT for other KETs

0 10 20 30 40 50 60 70 80 90 100

East Asia
Industrial Biotechnology
Nanotechnology

Photonics
Micro- and Nanoelectronics*

Advanced Materials
AMT for other KETs

0 10 20 30 40 50 60 70 80 90 100

* 2005 2006 average, change in unit values for 2002 2003 to 2005 2006.
Source: COMTRADE Database, NIW calculations

products with higher-than-average unit values, though

In AMT for other KETs, the composition of North
most of these products reported decreasing export
Americas exports by technology content is very
unit values compared with the same products in the
similar to the EU-28. This result indicates that both
competitor regions. In advanced manufacturing
regions specialise in trade in different products, each
technologies for other KETs, most East Asian export
region specialising in those products for which it has
products show an increasing technology content over
superior unit values. Like the EU-28, North
time. In advanced materials, East Asia reports a
Americas exports in micro- and nano-electronics
similar pattern of technology content as North
were focused on low technology content products, at
America: a mix of high and low technology content
least in the first half of the 2000s. In contrast to the
products. Photonics is clearly where East Asian
EU-28, almost all export products faced decreasing
exports focus on low technology content products, an
unit values compared with the export unit values of
indication of early stages in the value chains.
the same products in the competitor regions. In
advanced materials, North American exported Within the EU-28, export performance with respect to
products of varying technology content. technology content varies among Member States,
though the main patterns for the EU-28 can be
East Asia exports KET-related products which are
recognised for many of the largest Member States
classified mainly as low technology content products.
(Figure 5.7). In industrial biotechnology, a high share
The main exception is micro- and nano-electronics,
of exports with products showing a high and
where all products exported by East Asia have
increasing technology content can be found for
technology content. In addition, 46 % of
Denmark, the Czech Republic, Belgium, Austria,
nanotechnology exports in 2007 2011 were based on
France, Netherlands, the UK, Germany and Italy.

157
In nanotechnology, only the UK, Netherlands,
Belgium and Italy report a share above 50 % for
products with high and increasing technology content.
For photonics, the high share of EU-28 exports
products with high and increasing technology content
is due mainly to the export activities of France and
Sweden.
5.2.3. Type of competition and competitive
advantages in international trade
In addition to technology content, the type of
competition a countrys or regions products face in
international trade provides further, complementary
information on the position of the country/region in
international value chains. To simplify: product
market competition can be driven either by price or
by quality. Price competition dominates if the price
elasticity is high while at the same time product
differentiation (differentiating similar products by
quality characteristics such as durability, usability,
flexibility, additional performance characteristics) is
of little relevance. Price competition often indicates
that products are positioned earlier in the value chain,
while quality competition may be associated with
more complex products further along the values
chain.
Aiginger (1997, 2000) proposed a method to classify
products according to quality and price competition
based on the relation of a regions export unit values
to its import unit values on the one hand, and its trade
balance on the other. Products are price elastic (price
competition dominates) if export unit values which
are higher (lower) than import unit values lead to a
negative (positive) trade balance.
Conversely, products for which higher (lower) export
unit values than import unit values result in a positive
(negative) trade balance are price inelastic, in other
words quality competition dominates. For both types
of competition, a positive trade balance is an
indication that the region can build on a competitive
advantage for that type of competition. Combining
the relation of export unit values to import unit values
with the trade balance produces four quadrants
(Figure 5.8) in which a region can be positioned for
each KET-related product:
1. Quality competition with a quality advantage:
where export unit values exceed import unit
values and the trade balance is positive a
country or region can export more of a certain
product than it imports, despite higher prices.
2. Quality competition without quality advantage:
export unit values are lower than import unit
values while the trade balance is negative a
country or region imports more than it exports
despite lower prices, indicating that quality is
the main driver for trade.
3. Price competition with a price advantage: a
country or region shows lower export unit
158
values than import unit values and can translate
lower prices into a positive trade balance.
4. Price competition without a price advantage:
export unit values are higher than import unit
values in combination with a negative trade
balance.
As for technology content, the type of competition
that dominates the exports of country or region i in a
certain KET k is determined by the position p in the
quadrants shown in Figure 5.8 (p {1,2,3,4}) that
each individual product j belonging to KET area k
occupies, weighted by the products share in total
exports X of products related to KET area k of
country or region i. To calculate the type of
competition, the same data source is used as for
calculating the technology content of trade.
The results for the three main regions are reported in
Figure 5.9. Exports of KETs-related products by the
EU-28 face very different competition on
international markets. In advanced manufacturing
technologies for other KETs, most EU-28 exports
(64 %) concern products for which trade is
characterised by quality competition. For almost all
these products, the EU-28 has a quality advantage; in
other words, it is able to gain a positive trade balance
based on superior product quality. In nanotechnology,
industrial biotechnology and advanced materials, only
23 % to 34 % of EU-28 exports are based on quality
competition. Although the majority of EU-28 exports
in these KETs is characterised by price competition,
most of these exports benefit from price advantages.
This means that Member States specialise in those
price- sensitive products for which a cost-efficient
production in the EU is possible. In photonics and
micro-/nanoelectronics, most of the products exported
by the EU-28 are in price competition (89 % and 9 4%
respectively), and for the majority of these products
the EU has no price advantage.
North America reports a strong focus on exports
which face quality competition: it relies on a quality
advantage in international trade in the fields of
photonics (78 % of all exports in this KET) and
nanotechnology (54 %). In micro- and nano-
electronics, 41 % of North Americas exports fall into
this category, while 15 % are characterised by quality
competition, without having a quality advantage. In
the other three KETs, exports from North America
mainly face price competition, with a price advantage
over their main competitors.
East Asias trade in KET-related products is strongly
focused on price competition. In five KETs
industrial biotechnology, nanotechnology, micro- and
nanoelectronics, advanced materials and advanced
manufacturing technologies for other KETs East
Asia benefits from a price advantage, in other words a
cost-efficient production. Photonics is the only area
where East Asias exports are under major pressure,
as most of its products face price competition but
cannot compete on a price advantage. In each KET,
the share of KET-related products exported from East
Asia which are in markets dominated by quality
competition is lower than for North America, ranging
from 10 % (micro-and nanoelectronics) to 29 %
(photonics). The majority of these exports do not
have a quality advantage.
When examining the development of competition
types by KET over time, no clear trends for the
EU-28 emerge. In advanced manufacturing
based on price competition and price advantage
decreased. In photonics, the share of EU exports in
markets with price competition which could profit
from an EU price advantage has fallen substantially in
the last ten years, while the share of exports facing
price competition without a price advantage has
increased.
At the level of EU Member States (Figure 5.10), most
countries face price competition for the majority of
their KETs-related exports. Interestingly, for products
facing price competition some large Member States

Figure 5.7. Technology content of KET-related products by KET for selected Member States, 2007 2011 averages

high and increasing technology content high but decreasing technology content low but increasing technology content
low and decreasing technology content
Germany France United Kingdom
IB IB IB

NT NT NT

PH PH PH

ME* ME* ME*

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Italy Belgium Netherlands

IB IB IB

NT NT NT

PH PH PH

ME* ME* ME*

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Denmark Austria Sweden
IB IB IB

NT NT NT

PH PH PH

ME* ME* ME*

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Poland Czech Republic Hungary
IB IB IB

NT NT NT

PH PH PH

ME* ME* ME*

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Notes: IB: Industrial biotechnology; NT: Nanotechnology; PH: Photonics; ME: Micro and nanoelectronics; MA: Advanced materials;
MT: AMT for other KETs.
* 2005 2006 average, change in unit values for 2002 2003 to 2005 2006.
Source: COMTRADE Database, NIW calculations

technologies for other KETs, the share of EU exports (Germany, France, UK, Italy) and the Netherlands do
based on quality competition and quality advantage not appear to have any price advantage. By contrast,
increased during the 2000s, while the share of exports exports of price-sensitive KET-related products from

159
Sweden, Austria, Belgium, the Czech Republic and predominantly feature price disadvantages.
Hungary rely mostly on price advantages, though in
each Member State there are also some KETs with Quality competition dominates for only a few KETs
products that in each Member State. In Germany and Austria, most

Figure 5.8. Measuring the type of competition and competitive advantages of manufactured products

p=1: quality competition p=3: price competition UV ... unit value ($/kg)
and quality advantage and price advantage X ... exports
UVXijt > UVMijt UVXijt < UVMijt M ... imports
Q ... quantity (kg)
QXijt > QMijt QXijt > QMijt
... change over time
p=2: quality competition, p=4: price competition, i ... country/region (i w)
no quality advantage no price advantage w ... world (total trade)
UVXijt < UVMit UVXijt > UVMijt j ... KET-related product
QXijt < QMijt QXijt < QMijt k ... KET
t ... time
CP(p)ikt = j Xpij(k)t / ip Xpij(k)t CP ... Type of Competition
p ... position in quadrant

Source: NIW/ZEW based on Aiginger (1997)

Figure 5.9. Type of competition in trade with KET-related products, 2002 2011 averages
Quality competition and quality advantage Quality competition without quality advantage
Price competition and price advantage Price competition without price advantage
EU-28
Industrial Biotechnology
Nanotechnology
Photonics
Micro- and Nanoelectronics
Advanced Materials
AMT for other KETs

0 10 20 30 40 50 60 70 80 90 100
North America
Industrial Biotechnology
Nanotechnology
Photonics
Micro- and Nanoelectronics
Advanced Materials
AMT for other KETs

0 10 20 30 40 50 60 70 80 90 100
East Asia
Industrial Biotechnology
Nanotechnology
Photonics
Micro- and Nanoelectronics
Advanced Materials
AMT for other KETs

0 10 20 30 40 50 60 70 80 90 100

Source: COMTRADE Database. NIW calculations

exports in advanced manufacturing technologies for

160
other KETs rely on quality advantages and compete
on quality. In Denmark and Sweden, the same is true
for industrial biotechnology. Exports from the
Netherlands and Hungary in products related to
nanotechnology are also predominantly
markets are characterised by price competition, but
most Member States do not possess a price advantage
for these exports.

Figure 5.10. Type of competition in trade with KETs-related products for selected Member States, 2002 2011
averages

high and increasing technology content high but decreasing technology content low but increasing technology content
low and decreasing technology content not to be determined
Germany France United Kingdom
IB IB IB

NT NT NT

PH PH PH

ME ME ME

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Italy Belgium Netherlands

IB IB IB

NT NT NT

PH PH PH

ME ME ME

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Denmark Austria Sweden
IB IB IB

NT NT NT

PH PH PH

ME ME ME

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Poland Czech Republic Hungary
IB IB IB

NT NT NT

PH PH PH

ME ME ME

MA MA MA

MT MT MT

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

IB: Industrial biotechnology; NT: Nanotechnology; PH: Photonics; ME: Micro and nanoelectronics; MA: Advanced materials; MT: AMT
for other KETs
Source: COMTRADE Database, NIW calculations

based on a quality advantage in markets where they
face quality competition. In micro- and
nanoelectronics, Denmark is the only Member State
considered here which exports most of its products
based on quality competition and quality advantage.
In advanced materials, only the Netherlands is in the
same situation. In photonics, for all Member States
considered here apart from Poland, most export
5.2.4. Link between patenting and technology
content of products related to key
enabling technologies
The link between patenting activities and the
technology content of products provides another
indication of a countrys position in the value chains
of KET-related products. If the production of new
technological knowledge (as revealed through patent

161
applications) has a direct impact on the technology
content of traded products, one may conclude that
these products are closer to the technological frontier
and depend on a direct technology input from recent
efforts in developing new technology. In order to
examine this link, a countrys unit values of exports
of products based on a certain KET are regressed are
regressed on the patent activities of that country in the
same KET.
The level of export unit values (UV) for each product
j belonging to a KET k in country i in period t are
explained by the countrys patent activity in the
respective KET area k in a previous period t n. Since
unit values do not depend on country size, while
patent activity does, the latter is divided by country
population to derive a size-adjusted patent intensity
(PINT). Country-specific variables such as size
(GDP) and productivity (GDP per capita, PROD) are
used to control for the effects of market size and the
sophistication of the production system, while time
dummies are used to capture changes in prices over
time:
ln(UVX)ij,t = + 1 ln(PINT)ik,tn + 2 ln(GDP)i,t +
3 ln(PROD)i,t + t t dt + i with j k
The patent intensity indicates to what extent a country
produces new technological knowledge in a certain
subfield of KETs given the total resources available
in that country. All variables are measured in
logarithms. The model is estimated for each KET
separately as well as across all KETs for the period
2002 to 2011 for 39 countries (EU-28, US, Canada,
China, Japan, South Korea, Switzerland, Norway,
Iceland, Israel, Former Yugoslav Republic of
Macedonia, and Russia).
The estimation results confirm a positive link
between lagged patent intensity and unit values.
Across all six KETs, a 10 % increase in patenting
results in a 1.2 % increase in export unit values
(Figure 5.11). The impact of patenting on the
technology content of exports is largest in advanced
manufacturing technologies for other KETs (2.4 %
increase in unit values, or twice as high as for all
KETs) and micro-/nanoelectronics (2.3 % increase).
In industrial biotechnology, the elasticity of unit
values on patent intensity is 1.5 %, in nanotechnology
1.4 %, in advanced materials 1.0 % and in photonics
0.8 %.
The main findings also hold when only EU Member
States are considered. For the EU, the link between
patent intensity and unit values of exports is of
similar magnitude as for the entire set of countries. A
10 % increase of patent intensity would transfer into
an increase of export unit values of 1.0 %. For three
KETs, the link between patenting and technology
content of exports is stronger in the EU than for all 39
countries considered in this analysis. In micro-
/nanoelectronics, a 10 % increase in patent intensity in
162
the EU-28 results in a 2.7 % increase in export unit
values.
In photonics, the elasticity in the EU-28 is 1.0 % and
1.6 % in industrial biotechnology. In advanced
manufacturing technologies for other KETs, the EU
Member States report the same elasticity for patenting
as the total group of 39 countries. In nanotechnology
and advanced materials, the EU-28 elasticity of
patenting is somewhat lower (0.9 % and 0.8 %
respectively).
5.3. VALUE CHAIN ANALYSIS OF PROMISING KETS-
BASED PRODUCTS
5.3.1. Introduction
How do these general observations hold when
focusing on a number of product-specific value
chains? In this section, the value chains of two
promising products based on key enabling
technologies are analysed and discussed. First, the
selection of the two products, lipase enzymes and
accelerometers, is explained. A more detailed analysis
of the value chain of lipase enzymes and the
accelerometer is then provided.
The analysis begins with a detailed description of the
value chain, after which all relevant players in each
part of the value chain are identified, thereby
analysing the position of EU companies vis--vis
non-EU companies. The information and analyses are
based on expert interviews, articles, news sites and
market reports. The methodology used to analyse the
value chain is the same methodology as in the
feasibility study for an EU monitoring mechanism on
KETs (Van de Velde et al. 2013).
The analysis points out that EU firms play an
important role in essential parts of the value chain
even though the exact proportions of value added
captured by EU firms could not be retrieved.141
5.3.2. Selection of products
The economic importance and growth potential of the
candidate product has been the main selection
criterion. Furthermore, whether a candidate product
constitutes a relatively new application or is well-
established is another factor to consider. The value
chain analysis aims to focus on upcoming products
that are driven by technological innovation, and to
analyse how the EU performs in developing and
marketing new high technology products in the KETs
141
A major difficulty here is the possibility of estimating the
value added of a KETs-based product in the total product
range of a company. For example, in the case of foundries, no
information is disclosed on the share of the accelerometer
production versus total production. Moreover, this share tends
to change over time, particularly due to rapid shifts in market
demand. There is also often a problem of corporate
confidentiality to overcome.
area and how EU policies support this process. On the
basis of an extensive literature review, the enzyme
class lipases in industrial biotechnology and the
accelerometer in micro- and nanoelectronics were
chosen. The overall selection process is presented in
Figure 5.12.
Figure 5.11. Different steps in the selection of products
for the value chain analysis
Source: IDEA Consult
5.3.2.1. Lipase enzymes
The first key enabling technology selected is
industrial biotechnology because of its fundamental
role in the development of a bio-based economy. Bio-
based products are one of the six priority action lines
Figure 5.12. Link between patent output and unit values of exports, 2002 2011, change in unit values from a 10 %
change of patent output, by KET, 2002 2011
Source: COMTRADE and PATSTAT data, ZEW calculations
in the Communication A stronger European Industry
for Growth and Economic Recovery (European
Commission 2012 b). Within industrial
biotechnology, enzymes have been selected as the
product segment. Enzymes are one of the major
promising areas in industrial biotechnology. They
enable a broad range of applications and provide
several advantages over traditional chemistry,
including high selectivity, lower energy use and mild
reaction conditions. The market for enzymes has
grown rapidly over the past decade, and both
households and industry are becoming more
dependent on enzymatic catalysis. Even so, there
remains a vast untapped potential in the enzyme
market (Sarrouh et al. 2012). In 2010, the market for
industrial enzymes was worth USD 3.3 bn and is
expected to grow to USD 4.4 bn by 2015, a compound
163
annual growth rate of 6 % over those five years (BCC
Research 2011 a).
Within enzymes, lipases are, together with
carbohydrases, considered to have the highest growth
prospects (Global Industry Analysts 2012). Lipases
enzymes that catalyse chemical conversions of fats
have traditionally been used in detergents to remove
fat and oily stains. In addition, they are increasingly
used in the food industry, for instance in applications
involving dairy products and baking. In recent years,
lipases have also received more attention as highly
effective, versatile and flexible biocatalysts for
organic synthesis. This has opened up a whole new
range of possibilities, including the production of
basic chemicals, specialty chemicals,
pharmaceuticals, cosmetics and biodiesel. Lipases
have the potential to impact positively on several
industries, both in terms of competitiveness and
environmental friendliness (lower energy use, fewer
unwanted side-products). Therefore this product is
selected to analyse the position of the EU in the value
chain of this class of enzymes.
5.3.2.2. Accelerometer
Micro and nanoelectronics provide knowledge and
technologies which generate some 10 % of GDP. The
expected market size for this key enabling technology
is estimated to be USD 300 bn in 2015, with an
expected compound annual growth rate of 13 % (HLG
MNE 2011). It has enabled the rise of the information
age, impacts deeply on everyday life and is expected
to continue to do so. Given the increasing importance
of More-than-Moore (MtM) applications, a product
within MtM and more specifically in the segment of
micro-electro mechanical systems (MEMS) was
chosen. MEMS are products elementary for many
types of interactions of electronics with the outside
world, and provide a good example of the continued
Figure 5.13. Value chain decomposition for enzymes
Source: IDEA Consult
integration of digital and non-digital functions over
time (the MtM trend). This integration has enabled
MEMS to grow rapidly in recent years. The MtM
trend is a major evolution with potential for radical
innovations, and the relative weight of MtM in the
industry is expected to increase (ITRS 2010).
In the MEMS segment, rapid growth is expected for
inertial sensors (Yole Dveloppement 2012). One
example of fast-growing inertial sensors is the
accelerometer, which is the chosen product based on
this key enabling technology. It is a motion sensor
which measures the acceleration of a given object. It
was first introduced on a large scale in the automotive
sector but has since found its way into many
consumer electronics applications. In mobile devices
such as smartphones, accelerometers are key sensors
as they enable gesture recognition, user interface
control and activity monitoring. Other consumer
electronics applications include measuring motion in
gaming and sports applications. The accelerometer
has been an important sensor in the evolution towards
more intelligent consumer electronics devices
becoming increasingly aware of the environment and
the conditions of the device (Ryhnen 2010). As the
accelerometer is representative of two major
evolutions in micro- and nanoelectronics the trend
towards mobile, intelligent devices and the
integration of heterogeneous functions this product
has been selected.
5.4. VALUE CHAIN ANALYSIS OF LIPASE ENZYMES
5.4.1. Value chain decomposition
Figure 5.13 shows the value chain of lipase enzymes,
consisting of two broad phases. First there is the
selection and genetic engineering of an appropriate
microorganism capable of producing the enzyme of
interest. Then there is the actual production of the
enzyme. The latter phase can be subdivided into four
major steps. The first step is the development of the
production process, which entails discerning the right
conditions for fermentation (the following step) and
the up-scaling of production from laboratory scale to
commercial scale. Once the production process is
optimised, large-scale fermentation can occur. During
fermentation, the microorganisms grow on a substrate
and produce the enzyme of interest. Once
fermentation is complete, the next step is to separate
164
the enzymes from the fermentation mass (product
recovery). In the final step, the enzyme product is
purified. The necessity of this final step depends on
the application.
The value chain in Figure 5.13 applies not only to
lipase enzymes but to enzymes in general, and to a
large extent even to industrial biotechnology in
general.
5.4.2. EU activity along the value chain
In what follows, the key players in the lipase enzymes
value chain will be identified and discussed. First, the
lipase enzyme producers companies selling lipase
enzymes will be described. Next, companies that do
not sell lipases directly but contribute to the value
added of this product by executing a specific step in
the value chain will be discussed. The focus will be
on the companies providing services for the selection
and engineering of the microorganism, followed by
companies active in the second part of the value
chain, namely enzyme production. Table 5.1 provides
an overview of identified companies along the value
chain.
5.4.2.1. Lipase producers
Fifteen companies selling lipases have been
identified, almost half of them located in Europe. In
Denmark-based Novozymes, Europe hosts the
worlds largest player in the overall enzyme industry
with a market share of about 47 % of global enzyme
sales in 2011. With regard to lipase enzymes, the
company offers a broad portfolio of enzymes for
various applications, including as detergents and food
processing, but also more recent applications such as
biocatalysis for the pharmaceutical, cosmetics and
chemicals industry. Netherlands-based DSM is larger
than Novozymes in terms of revenue but has a lower
market share in enzymes (6 % compared with 47 % for
Novozymes 142). DSM is active particularly in lipases
for food applications and has recently underlined its
interest in this field by acquiring lipase technology
from US-based Verenium, primarily to extend its
activities in food applications. However, its lipase
portfolio is not as broad as that of Novozymes.
142
Source: Novozymes
The other European companies active in the lipase
segment are generally an order of magnitude smaller
than DSM and Novozymes. Many of them focus on a
number of specific applications. For example, AB
Enzymes (owned by the UK-based ABF) produces
lipases predominantly for food (baking) purposes,
while Biocatalysts focuses on dairy applications.
Eucodis Bioscience is a relatively young company
with a high share of lipases in its product portfolio,
targeting mainly pharmaceutical applications.
Germany-based C-lecta is also a relatively young
company, offering enzymes for a limited number of
applications. With its lipase products it focuses
notably on the production of specialty chemicals.
Outside the EU, the main emerging country is the US,
where the largest player is Dupont, due to its recent
acquisition of Genencor. Genencor is the second
largest enzyme producer in the world, with about half
the sales of Novozymes. While Genencor holds a
strong position in the food enzyme market, it has so
far not been able to play a leading role in the lipase
segment of the food market. In addition, unlike
Novozymes it is much less present in the emerging
markets for lipases, such as pharmaceutical and
chemical market applications. The other US
companies are significantly smaller than Genencor.
Codexis produces enzymes used to improve
production processes in the pharmaceutical industry
and is currently developing its lipase activities.
Verenium has achieved some success in the
commercial development of lipases, as illustrated by
the recent acquisition of its enzyme technology
including lipases by DSM. Dyadic is another rather
small company which owns a revolutionary
technology platform for the discovery and production
of enzymes, but it does not have a specific focus on
lipases.
Japan has two companies in Table 5.1. Amano
Enzymes has a long history as a specialty enzyme
developer, with key strengths in hydrolases. One of
its most successful products is a lipase which has
been widely produced in the past. However, Amano
does not serve the wide range of applications
Novozymes does. The other Japanese company,
Meito-Sangyo, is smaller than Amano but is also a
recognised player in the field of lipases, especially in
applications involving chiral transformations.
Two companies from India also appear in the list,
both offering a broad portfolio of lipases. This
reflects the emergence of India in the enzyme
industry as one of the countries with the highest level
of commitment to the development of industrial
biotech in Asia.
European companies are well represented in each
segment. In the detergent segment, Genencor and
Novozymes are the two companies with the highest
market shares. The food segment is dominated by
165
Novozymes and DSM. Novozymes is the clear
market leader in the biocatalysis segment.
5.4.2.2. Companies active in microorganism selection
and engineering
Lipase producers are not necessarily active in all parts
of the value chain. Other companies can be active in
specific segments. Only three companies have been
identified as providers of services for microorganism
selection and engineering services, the first part of the
value chain. This can be explained by the fact that the
first step of the value chain can be considered as a
core competence of most lipase producers. One
example of a company offering such services is DSM,
which provides guidance to other companies in the
development of their technologies. For example,
when a lipase producer is working on the
commercialisation of its products, DSM can propose
other microorganisms for expressing the gene of
interest in order to facilitate the up-scaling of enzyme
production. Similarly, Lonza possesses so-called
expression platforms in-house engineered
microorganisms enabling high enzyme production
levels.
5.4.2.3. Companies active in enzyme production
The large-scale production of enzymes requires
considerable investment in infrastructure, which often
creates a hurdle, especially for smaller (start-up)
companies. Outsourcing enzyme production can
deliver specific benefits to enzyme companies, not
only by reducing the financial risk, but also by
gaining access to the fermentation service providers
know-how. In this respect, it is important to note that
in industrial biotechnology, the scaling up of the
production of a given product (such as enzymes or
vitamins) from laboratory to commercial scale is
more difficult than in classical chemical production
processes (Wydra 2012). Therefore the fermentation
service providers often guide enzyme producers
through the gradual up-scaling of production of a new
enzyme, a service primarily used by smaller
companies, whereas large industrial players typically
organise the whole production in-house.
The companies identified as active in enzyme
production are shown in the lower part of Table 5.1.
These companies are known to be engaged in contract
production of enzymes for industrial purposes.
However, it should be noted that the importance of
lipase production in their production services may
vary over time, and is not fully disclosed by these
companies. This list therefore applies to industrial
enzyme production in general. Four companies are
located in Europe. DSM, by far the largest company
involved, is present also in the list of lipase
producers. Apart from producing enzymes under its
own brand name, it also offers a broad range of
fermentation services to other companies. Two other
companies, Eucodis Bioscience and Biocatalysts, are
Table 5.1. Overview of companies active along the lipase value chain
Total revenue 2011
Region Country Company name
(USD million)
Lipase producers
Netherlands DSM 9 048
Denmark Novozymes 1 891
a
UK ABF (AB Enzymes) 16 650 (127)
EU
Germany C-lecta n.a.
UK Biocatalysts Ltd n.a.
Austria Eucodis Bioscience n.a.
US Dupont (Genencor)a 38 000 (835b)
US Codexis 124
US Verenium 61
US Dyadic 10
Non-EU Japan Amano 1 074c
Japan Meito-Sangyo 176
India Advanced Enzymes 34
India Aumgene Bioscience n.a.
China Syncozymes n.a.
Companies active in microorganism selection and engineering
EU Netherlands DSM 9 048
Switzerland Lonza 2 019
Non-EU
India Aumgene Bioscience n.a.
Companies active in enzyme production
Netherlands DSM 9 048
EU Finland Galilaeus Oy. 1.9b
Austria Eucodis Bioscience n.a.
UK Biocatalysts Ltd n.a.
a
Switzerland Novartis (Sandoz) 58 566 (10 700)
Switzerland Lonza 2 019
Non-EU Israel Biodalia n.a.
Mexico Fermic n.a.
India Aumgene Bioscience n.a.
Source: IDEA Consult. Company turnover and main production sites are based on corporate annual reports and company website
information
Notes: a: in case the relevant activities are performed by a specific subsidiary, this subsidiary is listed in parentheses behind the parent
company. b: total revenue 2010; c: total revenue 2012; n.a. = not available

also identified as lipase producers. Finland-based
Galilaeus is a small company focusing on
fermentation services for various fields.
Outside the EU, Switzerland hosts two companies.
However, the presence of the large pharmaceutical
company Novartis is due to its ownership of Sandoz,
which offers a broad range of fermentation services to
its facilities in Germany, Austria and Italy. Lonza is
also a large company with a broad range of activities.
Given that the main fermentation site of this company
is located in the Czech Republic, it can be concluded
that many activities of interest here of the two Swiss
companies take place in Europe, particularly the
fermentation services accessible to European firms.
Fermic, based in Mexico, has an agreement with US-

166
based Verenium for the manufacturing of all of the
latters enzymes.
Table 5.2. Dominant companies per lipase application
field
Companies with highest
Application field
market shares
Detergent Genencor, Novozymes
Food DSM, Novozymes
Biocatalysis Novozymes
Source: IDEA Consult 143
The world enzyme market is dominated by a select
number of companies (Novozymes, Genencor, DSM).
In a market where product innovation is very
important, their extensive R&D capabilities allow
these companies to remain at the forefront. However,
a second important element in the enzyme industry is
the capability to produce enzymes on a large scale
(and therefore at a moderate cost). This is because the
scaling of manufacturing processes in industrial
biotech is far less straightforward than in classical
chemistry. Currently Novozymes, Genencor and
DSM (and to a lesser extent AB Enzymes) distinguish
themselves in the large-scale effective production of
enzymes. While many smaller firms are good at the
discovery of new enzymes with interesting properties,
the step to large-scale manufacturing is not easy to
take. As a consequence, the technology of the smaller
companies is often acquired by larger companies who
then set up large-scale production of the enzyme.
Europe is well-placed in this regard and should foster
its capabilities in large-scale enzyme production since
it gives an important competitive advantage.
5.4.3. EU position in the value chain of lipase
enzymes
With Novozymes (the worlds leading enzyme
supplier) and DSM, Europe has leading companies in
all lipase application fields. In addition, Europe hosts
a group of smaller companies which tend to specialise
in certain applications. There is considerable
competition from US companies, primarily Genencor.
However, this company only has a leading market
position in detergent lipases. Japan, on the other hand,
hosts two recognised players which are strong in
emerging applications such as pharmaceutical and
chemical applications. However, the lipase activities
of these companies do not have the scale of the large
EU players.
Looking at more specific parts of the value chain, a
significant share of fermentation services is provided
within Europe, especially when taking into account
several EU-based activities of companies with Swiss
143
Table 5.2. is based on replies from interviewees about the most
important players in each application field.
167
headquarters. No other major region emerges in this
segment. As for the first step of the value chain,
microorganism selection and engineering, only a few
companies were identified, one of which is based in
the EU. This represents a smaller segment though.
The analysis largely expert-driven confirms that
Europe is a key player in the global enzyme market
and holds a strong position in the subfield of lipases.
It has not been possible to calculate the value added
captured by European firms in the value chain of
lipases. In order to assess the performance of EU
companies, industry experts were asked to list the
companies with the highest market shares for each of
the three major application fields of lipases in order to
gain insight into which regions lead in the segment.
For companies not selling lipases but focusing on
specific parts of the value chain, information on the
relevance of lipases in their activities is typically
more difficult to find since these activities are more
remote from the end-product.
5.5. VALUE CHAIN ANALYSIS OF THE
ACCELEROMETER
5.5.1. Value chain decomposition
The accelerometer consists of two main functional
units: a mechanical component which senses the
acceleration (the sensor) and an electronic unit which
receives and translates the signals coming from the
mechanical component. The electronic unit is often
referred to as an application-specific integrated circuit
(ASIC), as its sole purpose is to receive and translate
signals from the mechanical component. Figure 5.14
is a schematic representation of the value chain of the
accelerometer. It covers two major phases, the design
and manufacturing of the ASIC and of the sensor. In
the design phase, a complete plan of the ASIC and
sensor is drawn, detailing all functional structures and
how they will be interconnected. These plans will be
translated into a format that can be used for
manufacturing. The manufacturing process can be
divided into four steps. The first step is the
fabrication of the ASIC and sensor on a large silicon
substrate (or wafer). The next step, wafer probing, is
to inspect the wafer for malfunctioning ASICs and
sensors. Then the two components are integrated into
one package, followed by a final phase of tests of the
accelerometer. Each step will be discussed in more
detail below.
The value chain of the accelerometer consists of
many steps. A company can opt to cover the whole
value chain itself or focus on a specific number of
steps. Companies covering the entire chain are
integrated device manufacturers (IDMs) and are
responsible for the design, manufacturing and sale of
their own products. However, other business models
exist as well. A company can focus on design but
leave manufacturing to another firm (a fabless
company). Companies focusing exclusively on
manufacturing chips commissioned by fabless
companies are referred to as foundries. Intermediate
forms can also exist: a company can manufacture part
of its products and outsource the rest to a foundry, or
be active only in the design phase by providing
design services to other companies.
This illustrates that the value added from creating an
accelerometer is divided among several companies,
each covering particular stages of the value chain. It
is therefore important to look not only at the end-
producers of this product but at all parts of the value
chain when assessing the competitiveness of the EU
in this product. In the following paragraphs, the key
players will be identified for each step of the value
chain. First the end-producers of accelerometers
(companies selling accelerometers) will be discussed,
followed by companies active in the first part of the
value chain (design) and then those active in the
second part of the value chain (manufacturing).
Figure 5.14. Value chain decomposition for the accelerometer
Source: IDEA Consult
5.5.1.1. Accelerometer producers
Table 5.3 lists the most important players in the
accelerometer industry. It is clear that Germany, the
US and Japan are the three countries covering the
majority of the market (in case of multinational
companies, the country assigned is the location of the
parent company).
Europe has a small group of large companies, while
in the US there is a larger group of somewhat smaller
companies. Japan has three major players on the
market, but two are the result of recent acquisitions
(Rohm Semiconductor acquired US-based Kionix,
while Murata Electronics acquired Finland-based
VTI). The widespread use of accelerometers started in
the automotive industry, which remains an important
market. All large companies in Table 5.3 have a
strong presence in the automotive sector. In this
respect it has been an advantage for companies in
Europe to have a strong domestic automotive market.
The interest of consumer electronics manufacturers in
accelerometers (and other MEMS) has grown
gradually as a result of their drive to give electronics
intelligent attributes. This posed a number of
technical challenges smaller chips, higher
production volumes and extreme cost consciousness.
168
STMicroelectronics was one of the first sensor
producers to spot the potential of the consumer
electronics market and develop large-scale production
facilities for that market. This has brought the
company solid growth and a strong position in this
segment. Robert Bosch is another key player in the
consumer electronics market. Strength in both the
automotive and consumer segments provides an
opportunity to operate at a large scale, thus reducing
costs. The main competition in the accelerometer
market comes from US-based companies. Analog
Devices and Freescale are two well-established
competitors, while Memsic and Invense are two
young, promising and innovative companies with
strong growth rates over the past years.
The accelerometer market was worth around
USD 1.5 bn in 2011 and dominated by companies
such as STMicroelectronics and Bosch (around 20 %
each), Freescale (around 10 %), Analog Devices,
Denso and VTI. The success of STMicroelectronics
and Bosch in both automotive and consumer
electronics makes them the two largest players in the
accelerometer market. The EU represents almost half
of the market in this segment, making it the leading
region. Two US companies, Freescale and Analog
Devices, also capture a significant share of the
market, although less than STMicroelectronics or
Bosch. Together with other US companies with
smaller market shares, the US has the second largest
market share after the EU. The third player, Japan,
has been able to capture a considerable share of the
market with Denso and the acquisition of VTI and
Kionix.
5.5.1.2. Companies active in design
Companies specialising in support for MEMS
producers in the design phase are often referred to as
design houses. Their task involves helping other
companies with the design and prototyping of a new
product. Design houses can help deliver a faster
time-to-market (which tends to be quite long for
MEMS in general) and ensure a good match of the
design with more conventional manufacturing
techniques.
Table 5.3 also lists companies active as design houses
for MEMS. It should be noted that the importance of
accelerometer design in the activities of these
companies varies over time and is not disclosed.
Therefore, the list applies to MEMS more generally.
Of the four identified actors, three are US-based.
France-based Movea is a young, innovative company
specialised in motion sensing. Design houses are a
relatively small segment (much smaller than the
foundry segment) since design is a core competence
of most accelerometer end-producers and is often
undertaken to a significant extent by these companies
themselves.
5.5.1.3. Companies active in manufacturing
Companies active in accelerometer manufacturing are
those which manufacture accelerometer products on
behalf of a second party. Most of the large companies
in Table 5.3 (Bosch, STMicroelectronics, Freescale,
Analog Devices) are IDMs covering the whole value
chain. However, this is not true for all accelerometer
producers. The US-based company Invensense, for
example, operates a fabless model by operating a
simplified and innovative manufacturing process.
Outsourcing is not unique to small companies such as
Invensense. Companies like Analog Devices also
outsource part of their MEMS manufacturing. The
main rationale behind outsourcing is cost reduction.
The manufacturing equipment for integrated circuits
is capital intensive and cost-competitive production is
only possible if done on a large scale. A high-volume,
quick-turnaround consumer segment can often be
better served by dedicated large-scale foundries. An
exception is the automotive market, where strict
compliance requirements are in place and production
is in most cases done internally. Another reason to
use foundry services is the expertise these companies
have in the successful up-scaling of production to
large volumes.
Table 5.3 lists the most important foundries active in
accelerometer manufacturing. This list consists of
companies that undertake manufacturing of
accelerometer products on behalf of a second party. It
should be noted that while the companies listed in the
table are known to have accelerometer production
capabilities, the importance of accelerometer
production in their total activities varies over time
and is not disclosed. Two types of foundries can be
distinguished: those making only MEMS products
and those active also in the regular electronics
markets (memory, microprocessors). The latter
category of firms is referred to as silicon foundries.
Foundries (whether MEMS-exclusive or not) often
develop their own technological (manufacturing)
competences and hold a number of patents on in-
house manufacturing technology. MEMS foundries
also tend to offer a number of services aimed at
facilitating the translation of design into a successful
169
product, going from co-design to custom-specific
process development and packaging and testing.
In Europe there is considerable foundry activity in
four countries. Silex Microsystems has grown
strongly in MEMS and has become a strong player
thanks to its in-house manufacturing technology
which allows close integration of the ASIC and the
sensor. However, the EU foundry companies are
relatively small; and as only a part of their revenues
stems from accelerometer production their weight is
much smaller than accelerometer producers such as
STMicroelectronics and Robert Bosch.
Outside the EU, US companies are somewhat less
present in the foundry segment than the end-product
market. Global Foundries is a leading silicon foundry,
but in MEMS it is currently a small player as it has
only recently moved into this field (including
accelerometers), attracted by the good market
prospects. Taiwanese companies have a stronger
presence on the foundry list than on the list of
accelerometer producers. The less prominent presence
of US companies and the strong emergence of
Taiwanese companies might be interrelated, as a
number of US companies employ Taiwanese
foundries for their manufacturing. For example,
Invensense operates as a fully fabless company, with
all the manufacturing done in Taiwan by TSMC.
In addition, Analog Devices also outsources the
manufacturing of the electronic component (ASIC) of
its accelerometer to TSMC.
The Taiwanese foundry giants TSMC and UMC are
two examples of the successful development of the
semiconductor industry in Taiwan, particularly of the
foundry segment. These firms make the large
majority of their sales in the mainstream
semiconductor segments but are increasingly used by
other companies for MEMS production because the
large foundries are able to produce at low cost and in
large volumes. The positive growth prospects of the
MEMS segment have caught their attention, and both
have been active in developing skills needed for
mechanical sensor production in recent years. Other
silicon foundries are also entering, or planning to
enter, the MEMS segment. Germany-based Xfab has
invested heavily in MEMS production capacities in
recent years and has recorded positive growth figures
in this segment. The competition from foundries in
the mainstream semiconductor industry is expected to
grow in the near future.
A possible future competitive threat for Europe is its
low level of investments over the past decade in
semiconductor production capacity. European
companies have followed a strategy of prolonging the
life of 150 mm and 200 mm fabs by using them for
MEMS production. Investments in 300 mm fabs have
been low compared to US and Asia. Europe currently
has a very limited market share in the mainstream
semiconductor segments, where production in most
cases is done on 300 mm wafers and with advanced from the US, Canada and Taiwan, where regular
technology. On the other hand, MEMS production silicon foundries as well as MEMS foundries are
can be achieved competitively on 150 mm or 200 mm increasingly used by producers.
wafer fabs and without using the latest technology.
EU companies perform well thanks to a strong
However, with the ongoing depreciation of the older
background in the automotive industry, good R&D
150 mm and 200 mm fabs and the expected move of
competence and particularly in the case of the two
the mainstream semiconductor industry to 450 mm
large IDMs a rapid understanding of the
technology, it is expected that within five to ten years
possibilities of new markets and the advantage of
all MEMS production will take place on 300 mm
large-scale production. However, an important future
wafer fabs. Given that there is currently little 300 mm
competitive threat exists as investments in new
production technology in Europe, there is a risk that
(300 mm wafer) production capacity have been low in
manufacturing of MEMS (and the associated value
Europe in recent years. Currently this is not a
added and employment) will increasingly move out of
problem for MEMS (accelerometer) production, but
Europe. Moreover, the increasing dependence on
as the mainstream semiconductor industry migrates to
foreign countries for enabling technologies such as
450 mm wafers, manufacturing of MEMS
micro- and nanoelectronics may at some point also
(accelerometers) will occur on 300 mm within five to
give rise to strategic concerns.
ten years. Therefore there is a risk that manufacturing
A study for the European Commission in 2012 of these products will move to other regions,
suggests that Europe needs to take advantage of the primarily Asia. For Europe it seems that the transition
shift to 450 mm production technology to catch up to 450 mm should be taken as an opportunity to
with its investment deficit in production capacity. safeguard its leading position in accelerometers, but
Indeed, once the new 450 mm technology is installed, also in MEMS as a whole.
significant spare 300 mm capacity will become
Again, in the case of the accelerometer it has not been
available in other regions, and it will not make
possible to calculate the value added captured by
economic sense for the EU to invest massively in
European firms in the value chain. For accelerometer
300 mm capacity. One of the proposed scenarios is to
producers however, market reports contain
install 450 mm capacity initially to safeguard the
information on the market share of each company,
current strong position in More than Moore
providing insights into how different regions are
(including MEMS and the accelerometer) and later to
performing in the accelerometer market. For
expand the scope to more advanced technology for
companies focusing on specific parts of the value
More Moore (mainstream semiconductor)
chain, information on the relevance of the
production. The investment costs will be so high that
accelerometer in their activities is typically more
they will need to be spread over several years, which
difficult to find since these activities are more remote
means that early commitment is needed.
from the end-product and is therefore not included in
5.5.2. The EU position in the value chain of the this chapter.
accelerometer
5.6. CONCLUSIONS AND POLICY IMPLICATIONS
The analysis shows that EU companies have a solid The results of the analysis show that Europe holds
position in the end-producers segment of the varying positions in the different KETs. Asia is
accelerometer market. The EU is represented by a gaining ground as a main producer of new
relatively small group of large companies that have a technological knowledge in KETs, thereby
strong base in the automotive market. These demonstrating fast market share gains.
companies have also been able to take significant
shares of the fast-growing consumer market. The Europe has a strong technological capacity, a
strongest competition in the end-producers segment substantial production base, is specialised in (mature)
comes from US-based companies which consist of a products with high technology content, but has to
mixture of well-established companies and some compete mainly on price. Moving to the higher end of
younger, more innovative companies. In the smaller the value chain is a real challenge.
design segment, only a few companies have been
identified, all of which are located in the EU or the
US. In the foundry segment the EU is also well
represented, with four companies active in four
different countries. Here the main competition comes

170
Table 5.3. Overview of companies active along the value chain of the accelerometer

Total revenue 2011

Region Country Company name
(USD million)
Accelerometer producers
Germany Robert Bosch 70 539
Netherlands STMicroelectronics 9 735
EU
Germany Infineon 5 479
a
France Sagem (Colibrys) 16 438 (16)
US Freescale Semiconductor 4 572
US Analog Devices 2 993
US Invensense 96
US Honeywell 36 529
Non-EU US MEMSIC 68
US Endevco n.a.
Japan Denso 37 660
a
Japan Murata Electronics (VTI) 7 130b (76c)
Japan Rohm Semiconductor (Kionix) 3 187 (n.a.)
Companies active in design
EU France Movea n.a.
US A.M. Fitzgerald n.a.
Non-EU US Nanoshift n.a.
US SVTC Technologies n.a.
Companies active in manufacturing
Sweden Silex Microsystems 47
France Tronics 15.2
EU
Germany Xfab n.a.
UK Semefab n.a.
Norway Sensonor 7.5
Israel TowerJazz 611
US Global Foundries 3 480
a
US Teledyne (Dalsa Semiconductor) 1 941 (212c)
US IMT 24
Non-EU Canada Micralyne 15
Taiwan Asia Pacific Microsystems n.a.
Taiwan TSMC 12 914
Taiwan UMC 3 855
Malaysia MEMSTech n.a.
a
US/Japan UT - SPP (Silicon Sensing Systems) n.a.
Source: IDEA Consult. Company turnover is based on corporate annual reports and company website information.
Notes: a= in case the relevant activities are performed by a specific subsidiary, this subsidiary is listed in parentheses behind the parent
company; b= total revenue 2012; c= total revenue 2010; n.a. = not available

171
Looking at its position in the production and trade of
KETs, Europe has both a strong technological
capacity and a substantial production base in all
KETs. Europe is, in contrast to emerging competitors
from East Asia, specialised in key enabling
technology products with high technology content.
However, most of these products seem to be mature
as they compete mostly on price, less on quality.
There are, however, differences between and within
KETs.
- Industrial biotechnology: high and increasing
technology content of exports, and a price
advantage. In industrial biotechnology, Europe
specialises in products with high technology
content, in other words products with higher
quality or further along the production chain.
During the past decade, Europe has been able to
further strengthen its position. Despite Europes
technological advance, exports predominantly
face price competition. However, Europe tends
to have a price advantage in international trade in
industrial biotechnology products which could
point to a more efficient production. Patenting is
a major driver for the technology content of
industrial biotechnology products.
- Nanotechnology: low and decreasing
technology content of exports but with a price
advantage. In the nanotechnology sector, the EU
position is less favourable. The technology
content of most products is lower than in the two
main competitor regions (North America and
East Asia) and going down, indicating a
specialisation in less complex products. EU
exports in nanotechnology compete mainly on
price and in most cases EU products enjoy a
price advantage. Patenting is important to
achieve high technology content but this effect is
less pronounced in Europe than in North
America and East Asia.
- Micro- and nanoelectronics: low and
decreasing technology content of exports with
no price advantage. The technology content of
products is low and decreasing, accompanied by
strong price competition for exports and no sign
of a price advantage for Europe. To maintain
high levels of technology content, patenting is
extremely important, having an even stronger
impact in Europe than in the other two regions.
Patent activities in Europe have not been
sufficient to improve its trade position, though.
- Photonics: high and increasing technology
content of exports and a price advantage. In
photonics, EU exports show high and increasing
levels of technology content, which primarily
face price competition. Most of its exports do not
show a price advantage relative to competitors.
The role of patenting is lower for the technology
172
content of photonics products than for most other
KETs. A conclusion could be that Europe is
specialised in high-end products in photonics,
while global markets are increasingly
characterised by price erosion.
- Advanced materials: high and increasing
technology content of exports and a price
advantage. Advanced materials are in a similar
position as industrial biotechnology products in
Europe. The technology content is high and
increasing for most of Europes exports.
International competition is driven by price
competition and Europe can build on price
advantages for most of its export products.
Patenting is of secondary relevance for the
technology content of products.
- Advanced manufacturing for other KETs:
Europe is leading; high technology content of
exports; a clear quality advantage. Advanced
manufacturing for other KETs is the KET area in
which Europe holds the most advanced position
in production and trade. The technology content
of exports is high, increasing and strongly based
on patenting. Most of Europes exports compete
on quality and Europe holds quality advantages
for most of its export products. Europe is the
leading region in this KET, which is also
confirmed by a high positive specialisation and a
high positive trade balance.
The EU position in each of the key enabling
technology value chains is summarised in Table 5.4.
A key challenge for European competitiveness policy
is to bring European industry onto a competitive
path that rests firmly on more innovative and
more complex products. In many KETs this would
mean a focus on more integrated technologies,
including those which link several KETs. Such a
product portfolio would imply a shift of competitive
pressure towards quality competition. In such an
environment, EU industry could better exploit its
competitive advantages and create real value on
several levels.
In order to achieve this, various approaches may be
followed:
- Improving the links between producers of basic
technological elements with producers of
components and final products;
- Strengthening cross-fertilisation of technology
developments across key enabling technologies;
- Fostering and reinforcing the development of
clusters along value chains in key enabling
technologies, including knowledge producers
(such as universities and research institutes) and
knowledge users.
Table 5.4. The position of the EU in the production and trade of KET-related products: summary overview
Industrial Nano- Micro-/nano- Photonics Advanced AMT for
biotechnology technology electronics materials other KETs
Technology content of high and low, mostly low and high and high and high and
exports increasing decreasing decreasing increasing increasing increasing
Type of competition mostly price price price price price mostly quality
competition, competition, competition, no competition, no competition, competition,
price advantage mostly with price advantage price advantage mostly with quality
price advantage price advantage advantage
Impact of patenting moderate low high low low high
Source: ZEW and NIW

The strategy of moving European industry to the
higher end of value chains could build on a strong
base in each basic element within each key enabling
technology. Policy should also consider the
advantages of global cooperation in the development
and deployment of new key enabling technologies
and new applications. This could mean cooperation
with specialised technology suppliers from other
world regions. On the other hand, successful
commercialisation of new applications often depends
on cooperating with the right customers in those
markets that set future trends (lead markets). It
could be more beneficial for European industry to
commercialise new key enabling technologies abroad
even if parts of the production will move to
dynamic markets abroad than to focus on European
markets with less promising long-term prospects.
Moving to the higher end of the value chain is at the
same time enormously complex and challenging.
While monitoring the developments in the entire
value chain, it is equally important to focus on
promising segments and the position of the EU in the
value chains of these segments. This is confirmed by
the analysis of the two promising KETs products
which have been analysed above.
Focusing on promising KET product segments: a
starting point for moving up the value chain? On
the basis of the analysis of the value chains of two
products, lipase enzymes (industrial biotechnology)
and the accelerometer (micro- and nanotechnology), a
qualitative assessment has been made of the strength
of the EU in these two selected value chains.
Although precise figures on the value added captured
by European companies are unavailable (due to
confidentiality issues and the lack of insight into the
share of a particular technology in the overall valued
added of a company) the results show that the EU is
a key player in the area of lipase enzymes and the
global enzyme market. Europe also has a solid
position in the end-producers segment of the
accelerometer, as some EU companies have taken a
significant share of the automotive and consumer
electronics markets, two markets where
accelerometers are applied. Competition is
nevertheless strong, especially in the foundry
segment.
It is interesting to note that although the general EU
position in the entire micro- and nanotechnology
value chain is weak, its position with respect to the
accelerometer value chain is good. This suggests that,
even if the overall position in a particular key
enabling technology is not optimal, there may always
be segments existing or emerging where the EU is
in a good position and where active policy support
can make a difference in the longer run. It is
important to observe and monitor these specific
segments closely, for instance through the future
KETs Observatory, and act in time in order to
stay ahead of the competition. A focused and
intensified policy in this respect might, in the long
run, lead to Europe moving up the global key
enabling technology value chains.

173
 REFERENCES
Aiginger, K. (1997), The use of unit values to discriminate between price and quality competition, Cambridge
Journal of Economics, 21, 571592
Aiginger, K. (2000), Europes position in quality competition, background report for the European
competitiveness report 2000, Working Paper, DG Enterprise and Industry, European Commission
Arora, A., Fosfuri, A. and Gambardella, A. (2002), Markets for technology: the economics of innovation and
corporate strategy, Cambridge: MIT Press
Ballhaus, W., Pagella, A., Vogel, C., Wilmsmeier, C. (2011), Faster, greener, smarter reaching beyond the
horizon in the world of semi-conductors. Published by Pricewaterhouse-Coopers AG Wirtschaftsprfungs-
gesellschaft
BCC Research (2011 a), Enzymes in industrial applications: global markets
BCC Research (2011 b), Smart materials and their applications: technologies and global markets
BCC Research (2012), Nanotechnology: a realistic market assessment
BCC Research (2013), Robotics: technologies and global markets
Beise, M. and Rammer, C. (2006), Local user-producer interaction in innovation and export performance of
firms, Small Business Economics, 27, 207222
Bleaney, M. and Wakelin, K. (2002), Efficiency, innovation and exports, Oxford Bulletin of Economics and
Statistics, 64, 315
Calleja, D. (2013), The role of KETs (Key Enabling Technologies) in the 2020 European innovation strategy,
The European Files, September 2013
Clarke, P. (2013), Processor market to surf tablet wave in 2013, EE Times, 21 January 2013
Deloitte (2013), 2013 global manufacturing competitiveness index. Deloitte Touche Tohmatsu and U.S.
Council on Competitiveness
Ebling, G. and Janz, N. (1999), Export and innovation activities in the German service sector: empirical
evidence at the firm level, ZEW Discussion Paper No. 9953, Mannheim: ZEW
ECSIP (2013), Study on the international market distortion in the area of KETs: a case analysis. Study for
Directorate-General for Enterprise and Industry. Copenhagen: ECSIP
European Commission (2010), European Competitiveness Report 2010. Luxembourg: Publications Office of
the European Union
European Commission (2012 a), A European strategy for key enabling technologies a bridge to growth and
jobs, COM(2012) 341, Brussels: European Commission
European Commission (2012 b), A stronger European industry for growth and economic recovery, COM(2012)
582, Brussels: European Commission
European Commission (2012 c), Innovating for sustainable growth: a bioeconomy for Europe, COM(2012) 60,
Brussels: European Commission
European Commission (2013 a), Public-private partnerships in Horizon 2020: a powerful tool to deliver on
innovation and growth in Europe, COM(2013) 494, Brussels: European Commission
European Commission (2013 b), Proposal for a Council Regulation on the bio-based industries joint
undertaking, COM(2013) 496, Brussels: European Commission
European Commission (2013 c), Monthly update on industrial policy indicators and analysis, Directorate-General
for Enterprise and Industry, July 2013
European Parliament (2013), Resolution of 2 July 2013 on innovating for sustainable growth: a bioeconomy for
Europe, Strasbourg: European Parliament
EPIA (2013), Global market outlook for photovoltaics 20132017. Brussels: European Photovoltaic Industry
Association
Festel (2010), Industrial biotechnology: market size, company types, business models, and growth strategies,
Industrial Biotechnology, 6(2), 8894

174
Gambardella, A., Giuri, P. and Luzzi, A. (2007), The market for patents in Europe, Research Policy, 36, 1163
1183
Global Industry Analysts (2012), Industrial enzymes A global strategic business report, March 2012
HLG KETs (2011), Final report, June 2011
HLG MNE (2011), Interim thematic report by the micro-/nanoelectronics sherpa team, High Level Group (HLG)
on Key Enabling Technologies
ICT MAN (2011), Exploring the potential of ICT components and systems manufacturing in Europe
ITRS (2010), More-than-Moore white paper, International Technology Roadmap for Semiconductors (ITRS)
J.P. Morgan Cazenove (2012), Electrical engineering and semiconductor equipment: Winners and losers in a
radically changing lighting market driven by LED
Kircher, M. (2012), How to turn industrial biotechnology into reality, New Biotechnology, 29(2), 243247
Lamoreaux, N. and Sokoloff, K. (1999), Inventive activity and the market for technology in the United States,
18401920, NBER Working Paper 7107, Washington D.C.: NBER
Lefebvre, E., Lefebvre, L.A. and Bourgault, M. (1998), R&D-related capabilities as determinants of export
performance, Small Business Economics, 10, 365377
MarketsandMarkets (2012), Quantum dots (QD) market global forecast & analysis (20122022), May 2012
OECD (2009), The bioeconomy to 2030: designing a policy agenda. Paris: OECD
Overton, G., Anderson, S.G., Belforte, D.A., Hausken, T. (2011), Skies may be clearing, but fog still lingers,
Laser Focus World, January 2011, 4042
Photonics21 (2012), Photonics technologies and markets for a low carbon economy, study prepared for the
Photonics Unit, DG CONNECT, European Commission
Roper, S. and Love, J.H. (2002), Innovation and export performance: evidence from UK and German
manufacturing plants, Research Policy, 32, 10871102
Ryhnen, T. (2010), Impact of silicon MEMS 30 years after, in Lindroos, V., Tilli, M., Lehto, A. and
Motooka, T. (eds.), Handbook of silicon based MEMS materials and technologies, pp. xvxxxii
Sarrouh, B., Marcos Santos, T., Miyoshi, A., Dias, R. and Azevedo, V. (2012), Up-to-date insight on industrial
enzymes applications and global market, Journal of Bioprocessing & Biotechniques, S4:002
Serrano, C.J. (2005), The market for intellectual property: evidence from the transfer of patents, mimeo,
University of Minnesota and Federal Reserve Bank of Minneapolis
SIA (2012), Semiconductor industry posts record-breaking revenues despite 2011 challenges, Semiconductor
Industry Association, 6 February 2012
Silicon Europe (2012), Silicon Europe: cluster alliance for European micro- and nano-electronics industry,
8 October 2012
Technology Strategy Board (2012), A landscape for the future of high value manufacturing in the UK.
Swindon: Technology Strategy Board
Van de Velde, E., Rammer, C., de Heide, M., Pinaud, F., Verbeek, A., Gehrke, B., Mertens, K., Debergh, P.,
Schliessler, P., van der Zee, F. and Butter, M. (2013), Feasibility study for an EU Monitoring Mechanism on
Key Enabling Technologies, Brussels: European Commission
Wagner, J. (1996), Export performance, human capital, and product innovation in Germany: a micro view,
Jahrbuch fr Wirtschaftswissenschaften, 47, 4045
Wakelin, K. (1998): Innovation and export behaviour at the firm level, Research Policy, 26, 829841
Wydra, S. (2012). Overview of policy measures of biobased products. Industrial biotechnology, 8(1), 812
Yole Dveloppement (2012), Status of the MEMS industry 2012

175
176
 Chapter 6.
STATISTICAL ANNEX
6.1. SECTORAL COMPETITIVENESS INDICATORS

6.1.1. Explanatory notes

Geographical coverage: all indicators refer to EU-27

Production index144: The production index is actually an index of final production in volume terms.
Labour productivity: this indicator is calculated by combining the indexes of production and number of persons
employed or number of hours worked145. Therefore, this indicator measures final production per person of final
production per hour worked.
Unit Labour Cost: it is calculated from the production index and the index of wages and salaries and measures
labour cost per unit of production. Wages and salaries is defined (Eurostat) as the total remuneration, in cash
or in kind, payable to all persons counted on the payroll (including homeworkers), in return for work done
during the accounting period, regardless of whether it is paid on the basis of working time, output or piecework
and whether it is paid regularly wages and salaries do not include social contributions payable by the employer.
Relative Trade Balance: it is calculated, for sector i, as (Xi-Mi)/(Xi+Mi), where Xi and Mi are EU-27 exports
and imports of products of sector i to and from the rest of the World.
Revealed Comparative Advantage (RCA): The RCA indicator for product i is defined as follows:

where: X=value of exports; the reference group (W) is the EU-27 plus 109 other countries (see list below); the
source used is the UN COMTRADE database. In the calculation of RCA, XEU stands for exports to the rest of the
world (excluding intra-EU trade) and XW measures exports to the rest of the world by the countries in the
reference group. The latter consists of the EU-27 plus the following countries: Albania, Algeria, Azerbaijan,
Argentina, Australia, Bahamas, Bahrain, Armenia, Bermuda, Bhutan, Bolivia (Plurinational State of), Bosnia
Herzegovina, Brazil, Belize, Belarus, Cambodia, Canada, Cape Verde, Central African Rep., Sri Lanka, Chile,
China, Colombia, Costa Rica, Croatia, Dominican Rep., Ecuador, El Salvador, Ethiopia, French Polynesia,
Georgia, Gambia, State of Palestine, Ghana, Greenland, Guatemala, Guyana, China, Hong Kong SAR, Iceland,
Indonesia, Iran, Israel, Cte d'Ivoire, Japan, Kazakhstan, Jordan, Rep. of Korea, Kyrgyzstan, Lebanon, China,
Macao SAR, Madagascar, Malawi, Malaysia, Maldives, Mauritania, Mauritius, Mexico, Other Asia, nes, Rep. of
Moldova, Montenegro, Montserrat, Namibia, Nepal, Aruba, New Caledonia, New Zealand, Nicaragua, Niger,
Nigeria, Norway, Pakistan, Panama, Paraguay, Peru, Qatar, Russian Federation, Rwanda, Saint Kitts and Nevis,
Saint Vincent and the Grenadines, Saudi Arabia, Senegal, Serbia, India, Singapore, Viet Nam, South Africa,
Zimbabwe, Suriname, Switzerland, Thailand, Togo, Tonga, United Arab Emirates, Tunisia, Turkey, Turks and
Caicos Isds, Uganda, Ukraine, Former Yugoslav Republic of Macedonia, Egypt, United Rep. of Tanzania, US,
Burkina Faso, Samoa, Yemen, Zambia.
Statistical nomenclatures: the indicators in Table 6.1 to Table 6.6 are presented at the level of divisions of the
statistical classification of economic activities in the European Community (NACE Rev.2146), while those in
Table 6.7 to Table 6.9.2 are presented in terms of divisions of the statistical classification of products by activity
(CPA).

144
The data are working-day adjusted for production.
145
The data are working-day adjusted for hours worked.
146
Compared to the statistical annexes of the previous publications, the new activity classification is used: NACE REV 2. The
correspondence tables from NACE Rev. 2 NACE Rev. 1.1 and from NACE Rev. 1.1 to NACE Rev. 2, are available on:
http://epp.eurostat.ec.europa.eu/portal/page/portal/nace_rev2/introduction

177
Table 6.10 uses extended balance of payments services classification. In terms of data sources: Table 6.1 to
Table 6.6 are based on Eurostats short-term indicators data. Table 6.7, Table 6.8, Table 6.9.1 and Table 6.9.2
are based on United Nations COMTRADE. Table 6.10 is based on IMF balance of Payments. Royalties and
license fees were not included as it is not related to a special service activity.

178
Table 6.1. EU-27 - Industry production index, annual growth rate (%)
Code Average
Sector 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
(NACE Rev. 2) 2007-2012
B MINING AND QUARRYING -2.8 0.5 -2.8 -1.9 -6.0 -3.8 -0.1 -3.6 -10.7 -0.3 -7.0 -5.8 -5.5

C MANUFACTURING 0.1 -0.8 0.3 2.5 1.6 4.8 4.2 -1.9 -15.3 7.3 4.5 -2.2 -1.8

C10 Manufacture of food products 1.4 1.9 0.5 2.1 2.4 1.3 1.9 -0.5 -1.0 2.2 1.0 -0.7 0.2
C11 Manufacture of beverages 2.7 1.6 1.3 -2.3 0.9 3.7 1.2 -2.1 -3.2 -1.0 6.5 -2.8 -0.6
C12 Manufacture of tobacco products -1.6 -2.1 -5.8 -11.4 -5.5 -4.7 1.5 -11.7 -0.7 -5.9 -6.2 -4.5 -5.9
C13 Manufacture of textiles -2.9 -4.5 -3.4 -4.8 -5.9 -0.8 -1.1 -10.2 -18.0 7.8 -2.1 -6.0 -6.1
C14 Manufacture of wearing apparel -4.7 -11.5 -7.3 -5.6 -10.4 -0.5 -0.6 -7.6 -14.0 -1.1 -4.1 -5.8 -6.6
C15 Manufacture of leather and related products -5.7 -8.3 -6.8 -10.2 -9.0 -2.9 -5.7 -8.1 -14.1 2.1 5.4 -4.2 -4.0
Manufacture of wood and of products of wood and cork,
C16 except furniture; manufacture of articles of straw and -4.2 0.7 2.3 3.2 0.2 4.2 0.7 -9.1 -14.9 2.9 2.6 -4.3 -4.8
plaiting materials
C17 Manufacture of paper and paper products -2.0 3.4 1.4 2.9 -0.1 3.9 2.6 -3.2 -8.7 6.0 -1.1 -1.5 -1.8
C18 Printing and reproduction of recorded media -1.9 -0.6 -1.2 1.4 2.4 0.2 0.7 -2.2 -7.9 -0.3 -0.5 -6.6 -3.6
C19 Manufacture of coke and refined petroleum products -0.9 0.9 1.3 4.6 0.7 -0.9 0.4 1.0 -8.0 -2.0 -1.7 -1.9 -2.6
C20 Manufacture of chemicals and chemical products -1.9 1.9 -0.1 3.3 2.0 3.4 2.9 -3.3 -12.5 10.3 1.3 -1.6 -1.4
Manufacture of basic pharmaceutical products and
C21 10.9 8.6 5.0 -0.2 4.7 5.9 0.4 0.6 2.9 4.9 1.6 0.2 2.0
pharmaceutical preparations
C22 Manufacture of rubber and plastic products -0.3 0.0 1.9 1.8 0.9 3.9 4.5 -4.6 -13.9 7.3 3.8 -3.4 -2.4
C23 Manufacture of other non-metallic mineral products -0.6 -1.7 0.3 1.7 0.5 4.2 1.8 -6.7 -19.4 1.9 3.2 -8.4 -6.2
C24 Manufacture of basic metals -1.2 -0.4 0.2 4.9 -0.7 6.3 1.3 -3.6 -27.1 18.1 3.6 -5.1 -4.0
Manufacture of fabricated metal products, except machinery
C25 0.4 -0.6 0.9 2.6 1.5 4.8 6.1 -3.0 -22.6 6.9 7.2 -3.3 -3.6
and equipment
C26 Manufacture of computer, electronic and optical products -6.3 -10.2 0.5 6.4 2.8 9.0 7.6 0.7 -17.2 7.2 5.1 -2.0 -1.6
C27 Manufacture of electrical equipment -0.7 -4.1 -1.6 2.3 0.9 8.4 4.3 -0.7 -21.0 11.4 4.3 -2.3 -2.3
C28 Manufacture of machinery and equipment n.e.c. 1.3 -1.8 -0.8 4.1 3.9 8.4 8.4 1.4 -26.8 10.6 11.5 0.5 -1.7
C29 Manufacture of motor vehicles, trailers and semi-trailers 1.6 0.7 1.6 4.5 1.4 3.3 6.1 -5.9 -25.1 21.6 12.1 -3.1 -1.4
C30 Manufacture of other transport equipment 1.2 -3.5 0.5 0.4 2.0 7.8 4.8 3.9 -5.7 -0.7 4.1 2.8 0.8
C31 Manufacture of furniture -2.3 -5.1 -2.5 0.3 1.1 3.8 3.3 -5.0 -16.7 -1.0 2.0 -5.6 -5.5
C32 Other manufacturing 3.6 3.0 -2.2 1.2 0.9 5.3 1.7 -1.5 -6.9 8.2 3.1 0.0 0.5
C33 Repair and installation of machinery and equipment 0.3 -4.0 -2.0 4.5 1.1 7.9 4.5 3.7 -10.2 2.2 4.5 -1.5 -0.4
ELECTRICITY, GAS, STEAM AND AIR CONDITIONING
D 2.0 0.9 3.0 2.3 2.0 0.9 -0.7 -0.1 -4.5 4.1 -4.1 0.3 -0.9
SUPPLY
WATER SUPPLY; SEWERAGE, WASTE MANAGEMENT
E N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
AND REMEDIATION ACTIVITIES

F CONSTRUCTION 1.1 0.3 1.7 0.7 2.8 3.4 2.7 -2.9 -7.5 -4.7 0.2 -5.2 -4.1

N/A: Data not available, Source: Eurostat

179
Table 6.2. EU-27 - Number of persons employed, annual growth rate (%)
Code Average
(NACE Rev. 2)
Sector 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
2007-2012
B MINING AND QUARRYING -3.4 -4.6 -4.5 -4.6 -3.1 -3.8 -3.5 -1.5 -3.8 -4.1 -3.4 -1.4 -2.9
C MANUFACTURING 0.0 -1.9 -2.0 -1.9 -1.3 -0.8 0.5 -0.3 -7.3 -3.7 0.6 -0.3 -2.2
C10 Manufacture of food products -0.6 -0.9 -0.5 -1.2 0.1 -0.2 0.0 -0.1 -1.9 -0.5 0.5 -0.3 -0.5
C11 Manufacture of beverages -1.8 -1.1 -1.8 -1.2 -1.3 -1.4 -0.1 -1.2 -6.4 -1.9 -1.5 -1.6 -2.5
C12 Manufacture of tobacco products -4.2 -0.3 -5.0 -5.2 -2.1 0.1 -11.1 -9.9 -6.0 -6.8 -2.4 -4.9 -6.0
C13 Manufacture of textiles -3.2 -5.0 -7.2 -6.3 -4.3 -5.9 -5.3 -6.3 -12.9 -5.9 -2.7 -2.4 -6.1
C14 Manufacture of wearing apparel -3.1 -3.6 -3.9 -6.2 -7.7 -5.8 -5.9 -6.7 -13.1 -8.5 -1.8 -2.7 -6.6
C15 Manufacture of leather and related products -1.2 -0.6 -4.4 -6.7 -5.6 -2.7 -3.7 -5.9 -12.5 -3.6 4.1 0.4 -3.7
Manufacture of wood and of products of wood and cork,
C16 except furniture; manufacture of articles of straw and -0.9 -1.5 -1.3 -1.4 -0.6 -1.3 0.6 -2.3 -12.6 -3.4 -0.1 -3.2 -4.4
plaiting materials
C17 Manufacture of paper and paper products -1.7 -0.7 -3.0 -1.6 -2.6 -2.6 -2.8 -2.2 -5.5 -2.3 -0.5 -1.9 -2.5
C18 Printing and reproduction of recorded media -0.2 -2.1 -4.0 -1.9 -3.3 -1.6 -0.1 -2.3 -7.0 -4.8 -3.6 -3.6 -4.3
C19 Manufacture of coke and refined petroleum products -1.9 -3.0 -3.2 -1.7 -2.9 -3.1 0.9 -0.7 -3.3 -2.8 -2.8 0.7 -1.8
C20 Manufacture of chemicals and chemical products -0.8 -1.6 -2.6 -3.3 -2.2 -1.2 -0.6 -2.3 -4.5 -2.3 -0.2 -0.1 -1.9
Manufacture of basic pharmaceutical products and
C21 1.8 2.2 -0.5 -2.4 -1.4 1.6 0.3 -2.5 -3.6 -1.0 -0.4 1.2 -1.3
pharmaceutical preparations
C22 Manufacture of rubber and plastic products 1.0 -0.9 0.2 0.0 -0.8 -0.8 1.6 0.6 -7.0 -2.6 1.3 0.6 -1.5
C23 Manufacture of other non-metallic mineral products -0.6 -2.3 -2.8 -2.1 -1.0 -0.7 1.2 -2.1 -10.7 -6.5 -1.9 -3.0 -4.9
C24 Manufacture of basic metals -0.2 -4.0 -2.9 -4.2 -1.1 -1.0 -0.6 -0.6 -8.4 -5.4 1.0 -1.4 -3.0
Manufacture of fabricated metal products, except machinery
C25 1.0 -1.0 -1.1 0.2 -0.2 1.4 3.3 2.6 -8.4 -5.4 1.6 0.5 -1.9
and equipment
C26 Manufacture of computer, electronic and optical products 1.9 -5.6 -4.4 -2.9 -1.3 -0.8 1.2 -1.8 -8.7 -3.9 1.2 -1.9 -3.1
C27 Manufacture of electrical equipment 0.5 -3.9 -4.1 -1.4 -0.6 0.9 2.5 1.1 -8.2 -2.0 3.3 0.5 -1.1
C28 Manufacture of machinery and equipment n.e.c. 1.0 -1.5 -2.2 -2.4 -0.9 0.7 3.0 1.9 -5.9 -5.0 2.7 1.9 -0.9
C29 Manufacture of motor vehicles, trailers and semi-trailers 1.7 -1.0 -0.4 0.2 -0.8 -1.0 -0.2 0.8 -8.9 -2.8 2.9 1.3 -1.4
C30 Manufacture of other transport equipment 0.0 -1.7 -2.7 -1.7 0.3 0.8 2.7 1.7 -1.5 -4.7 -0.3 1.0 -0.8
C31 Manufacture of furniture 0.5 -3.4 0.2 -2.6 -2.5 -1.3 0.3 -2.1 -9.6 -8.4 -1.7 -3.1 -5.0
C32 Other manufacturing 1.1 -1.6 -0.2 -1.0 -1.7 -0.5 0.3 -0.1 -3.1 -1.9 -0.9 0.8 -1.1
C33 Repair and installation of machinery and equipment 0.2 -2.3 -2.0 -0.5 -0.5 0.5 0.6 3.5 -1.8 -2.8 -1.4 1.0 -0.3
ELECTRICITY, GAS, STEAM AND AIR
D -2.7 -5.1 -4.9 -3.6 -2.4 -1.2 -1.6 -0.9 1.8 -0.2 0.4 -1.8 -0.1
CONDITIONING SUPPLY
WATER SUPPLY; SEWERAGE, WASTE
E -0.4 -0.3 0.5 -1.0 -1.8 1.9 0.7 -1.0 -0.3 -0.3 -0.3 1.1 -0.2
MANAGEMENT AND REMEDIATION ACTIVITIES
F CONSTRUCTION 1.9 0.6 1.7 2.1 3.7 3.8 4.8 -0.3 -7.1 -5.5 -3.0 -3.9 -4.0
N/A: Data not available, Source: Eurostat

180
Table 6.3. EU-27 - Number of hours worked, annual growth rate (%)
Code Average
Sector 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
(NACE Rev. 2) 2007-2012
B MINING AND QUARRYING -2.7 -4.7 -5.3 -3.8 -3.0 -4.8 -3.6 -1.3 -5.4 -2.5 -2.3 -0.1 -2.3
C MANUFACTURING -1.2 -2.4 -2.5 -1.4 -1.5 -0.1 0.1 -0.7 -9.6 -0.7 1.5 -0.7 -2.1
C10 Manufacture of food products -1.1 -1.8 -2.0 -0.3 -0.5 -0.2 -0.5 0.2 -2.6 0.3 0.2 -0.7 -0.5
C11 Manufacture of beverages -0.2 -3.5 -0.3 0.5 -2.9 -4.0 -1.2 -1.7 -4.6 -4.5 -0.3 -1.4 -2.5
C12 Manufacture of tobacco products 2.0 -2.3 -10.3 -1.9 -3.8 -8.7 -2.5 -10.6 -6.1 -4.1 -1.3 -1.2 -4.7
C13 Manufacture of textiles -4.4 -5.6 -7.4 -7.2 -5.4 -5.7 -3.2 -5.7 -15.3 -0.6 -0.3 -1.4 -4.8
C14 Manufacture of wearing apparel -3.5 -2.3 -3.1 -3.7 -4.1 -3.7 -5.8 -6.9 -15.3 -8.3 0.3 -3.1 -6.8
C15 Manufacture of leather and related products -1.9 -0.6 -1.8 -2.6 -4.1 -0.9 -4.7 -7.1 -12.0 -0.6 3.9 0.4 -3.2
Manufacture of wood and of products of wood and cork,
C16 except furniture; manufacture of articles of straw and -3.3 -1.2 -1.9 -1.1 -2.2 -0.4 -0.3 -3.0 -13.6 0.1 0.2 -2.1 -3.8
plaiting materials
C17 Manufacture of paper and paper products -1.8 -0.4 -2.9 -1.9 -2.0 -0.9 -1.3 -4.1 -7.7 -0.5 0.2 -1.6 -2.8
C18 Printing and reproduction of recorded media -1.2 -3.5 -4.4 -2.9 -2.3 -0.1 0.2 -2.0 -6.1 -3.7 -2.4 -5.1 -3.9
C19 Manufacture of coke and refined petroleum products -1.9 -3.9 -2.1 -0.4 0.0 -3.7 0.4 2.0 -9.0 -2.2 -1.8 1.0 -2.1
C20 Manufacture of chemicals and chemical products -2.3 -2.2 -2.7 -2.0 -3.2 -0.9 -1.6 -1.9 -5.5 -1.5 0.8 0.8 -1.5
Manufacture of basic pharmaceutical products and
C21 0.4 1.9 -0.2 -0.7 -1.6 0.0 0.9 0.0 -1.9 -0.8 -0.1 1.8 -0.2
pharmaceutical preparations
C22 Manufacture of rubber and plastic products -0.3 -1.8 -1.5 -0.2 -1.4 1.6 0.6 -0.4 -9.4 0.9 2.2 0.3 -1.4
C23 Manufacture of other non-metallic mineral products -2.6 -3.3 -3.2 -1.3 -1.0 -0.5 0.4 -2.7 -12.4 -2.4 -0.4 -3.5 -4.4
C24 Manufacture of basic metals -1.6 -3.2 -4.7 -2.1 -2.1 0.0 -0.5 -0.9 -13.1 1.8 2.6 -2.1 -2.5
Manufacture of fabricated metal products, except machinery
C25 -0.5 -1.3 -2.0 -0.3 -1.0 1.7 2.2 2.9 -11.7 -0.7 1.9 0.4 -1.6
and equipment
C26 Manufacture of computer, electronic and optical products -0.3 -5.0 -4.4 -2.9 -1.5 -0.7 0.3 -1.0 -12.2 -2.4 -0.1 -1.7 -3.6
C27 Manufacture of electrical equipment -1.2 -2.9 -3.8 -1.6 -1.9 2.5 1.7 0.6 -13.4 3.4 3.3 -1.2 -1.7
C28 Manufacture of machinery and equipment n.e.c. -0.6 -2.3 -2.2 -1.2 -1.3 1.5 2.5 1.7 -11.2 -0.2 4.3 0.7 -1.1
C29 Manufacture of motor vehicles, trailers and semi-trailers 0.4 -1.8 -0.9 0.4 -0.3 -0.6 0.8 -1.4 -14.2 4.1 4.5 -0.2 -1.7
C30 Manufacture of other transport equipment -1.4 -2.0 -1.8 -2.3 -0.4 1.0 0.9 1.3 -1.9 -3.8 -0.4 2.1 -0.6
C31 Manufacture of furniture 0.7 -4.6 -3.3 -0.5 -3.4 0.8 0.3 -2.9 -11.9 -5.5 -0.8 -2.8 -4.9
C32 Other manufacturing 0.2 -3.2 -2.5 0.3 -2.8 -0.8 0.6 0.4 -5.7 0.2 2.4 1.5 -0.3
C33 Repair and installation of machinery and equipment -1.5 -3.0 -3.4 -2.7 0.0 0.9 0.6 1.6 0.1 -3.7 -0.1 0.6 -0.3
ELECTRICITY, GAS, STEAM AND AIR
D -1.8 -4.9 -4.6 -2.0 -0.3 -1.4 -1.0 -0.1 -0.9 -0.5 1.2 -2.7 -0.6
CONDITIONING SUPPLY
WATER SUPPLY; SEWERAGE, WASTE
E -1.2 -1.4 -0.9 1.2 -2.2 -0.2 0.2 0.7 -2.8 0.5 0.3 1.6 0.0
MANAGEMENT AND REMEDIATION ACTIVITIES
F CONSTRUCTION 0.6 -1.7 0.7 1.0 7.2 2.9 2.7 -1.3 -9.3 -7.2 -0.8 -1.7 -4.1
N/A: Data not available, Source: Eurostat

181
Table 6.4. EU-27 - Labour productivity per person employed, annual growth rate (%)
Code Average
Sector 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
(NACE Rev. 2) 2007-2012
B MINING AND QUARRYING 0.6 5.4 1.8 2.8 -3.0 0.0 3.5 -2.1 -7.1 4.0 -3.7 -4.4 -2.7
C MANUFACTURING 0.1 1.2 2.3 4.5 3.0 5.6 3.7 -1.6 -8.7 11.4 3.9 -1.9 0.4

C10 Manufacture of food products 2.0 2.8 1.0 3.3 2.3 1.5 1.9 -0.4 0.9 2.7 0.5 -0.4 0.7
C11 Manufacture of beverages 4.6 2.7 3.2 -1.2 2.3 5.2 1.3 -0.9 3.5 1.0 8.1 -1.3 2.0
C12 Manufacture of tobacco products 2.7 -1.8 -0.8 -6.6 -3.5 -4.8 14.1 -2.0 5.7 1.0 -3.9 0.4 0.2
C13 Manufacture of textiles 0.3 0.5 4.1 1.6 -1.7 5.4 4.5 -4.1 -5.8 14.5 0.6 -3.7 0.0
C14 Manufacture of wearing apparel -1.7 -8.2 -3.5 0.6 -2.9 5.7 5.6 -1.0 -1.0 8.1 -2.4 -3.2 0.0
C15 Manufacture of leather and related products -4.5 -7.8 -2.5 -3.7 -3.7 -0.2 -2.1 -2.3 -1.8 6.0 1.2 -4.6 -0.4
Manufacture of wood and of products of wood and cork,
C16 except furniture; manufacture of articles of straw and -3.3 2.2 3.7 4.7 0.8 5.6 0.1 -7.0 -2.7 6.5 2.7 -1.2 -0.4
plaiting materials
C17 Manufacture of paper and paper products -0.3 4.1 4.5 4.6 2.6 6.6 5.5 -1.0 -3.4 8.5 -0.6 0.4 0.7
C18 Printing and reproduction of recorded media -1.8 1.5 2.9 3.3 5.8 1.9 0.8 0.1 -0.9 4.7 3.2 -3.1 0.7
C19 Manufacture of coke and refined petroleum products 1.0 4.0 4.7 6.4 3.7 2.2 -0.5 1.7 -4.9 0.8 1.1 -2.6 -0.8
C20 Manufacture of chemicals and chemical products -1.1 3.6 2.6 6.8 4.2 4.7 3.5 -1.1 -8.4 12.9 1.5 -1.5 0.5
Manufacture of basic pharmaceutical products and
C21 8.9 6.3 5.5 2.3 6.2 4.2 0.1 3.2 6.7 6.0 2.0 -1.0 3.3
pharmaceutical preparations
C22 Manufacture of rubber and plastic products -1.3 0.9 1.7 1.8 1.7 4.7 2.9 -5.2 -7.4 10.1 2.5 -4.0 -1.0
C23 Manufacture of other non-metallic mineral products 0.0 0.6 3.2 3.9 1.5 4.9 0.6 -4.7 -9.7 9.0 5.2 -5.6 -1.4
C24 Manufacture of basic metals -1.0 3.7 3.2 9.5 0.4 7.4 1.9 -3.0 -20.5 24.8 2.6 -3.8 -1.0
Manufacture of fabricated metal products, except machinery
C25 -0.6 0.4 2.0 2.4 1.7 3.4 2.7 -5.5 -15.5 13.0 5.5 -3.8 -1.7
and equipment
C26 Manufacture of computer, electronic and optical products -8.1 -4.8 5.2 9.6 4.2 9.8 6.3 2.6 -9.3 11.5 3.8 -0.1 1.5
C27 Manufacture of electrical equipment -1.2 -0.2 2.6 3.7 1.5 7.4 1.8 -1.8 -14.0 13.7 1.0 -2.8 -1.2
C28 Manufacture of machinery and equipment n.e.c. 0.2 -0.3 1.5 6.7 4.8 7.6 5.3 -0.5 -22.2 16.4 8.5 -1.4 -0.7
C29 Manufacture of motor vehicles, trailers and semi-trailers -0.1 1.7 2.0 4.3 2.2 4.3 6.4 -6.7 -17.8 25.1 8.9 -4.4 0.0
C30 Manufacture of other transport equipment 1.2 -1.8 3.3 2.1 1.7 7.0 2.0 2.2 -4.3 4.2 4.4 1.7 1.6
C31 Manufacture of furniture -2.7 -1.7 -2.6 3.0 3.7 5.1 3.0 -3.0 -7.8 8.1 3.7 -2.5 -0.5
C32 Other manufacturing 2.4 4.7 -2.0 2.2 2.7 5.8 1.4 -1.4 -3.9 10.3 4.0 -0.8 1.5
C33 Repair and installation of machinery and equipment 0.1 -1.7 0.0 5.0 1.6 7.4 3.8 0.2 -8.6 5.1 5.9 -2.4 -0.1
ELECTRICITY, GAS, STEAM AND AIR
D 4.8 6.3 8.3 6.1 4.5 2.1 0.9 0.8 -6.2 4.3 -4.5 2.1 -0.8
CONDITIONING SUPPLY
WATER SUPPLY; SEWERAGE, WASTE
E N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
MANAGEMENT AND REMEDIATION ACTIVITIES

F CONSTRUCTION -0.8 -0.3 0.0 -1.4 -0.9 -0.4 -2.0 -2.6 -0.4 0.8 3.3 -1.4 -0.1

N/A: Data not available, Source: Eurostat

182
Table 6.5. EU-27 - Labour productivity per hour worked, annual growth rate (%)
Code Average
Sector 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
(NACE Rev. 2) 2007-2012
B MINING AND QUARRYING -0.1 5.4 2.6 2.0 -3.1 1.0 3.7 -2.3 -5.6 2.2 -4.8 -5.7 -3.3
C MANUFACTURING 1.3 1.6 2.9 3.9 3.2 4.9 4.1 -1.2 -6.3 8.0 3.0 -1.5 0.3
C10 Manufacture of food products 2.5 3.8 2.6 2.4 2.9 1.5 2.4 -0.7 1.6 1.9 0.8 0.0 0.7
C11 Manufacture of beverages 2.9 5.3 1.6 -2.8 3.9 8.0 2.4 -0.4 1.5 3.7 6.8 -1.4 2.0
C12 Manufacture of tobacco products -3.6 0.2 5.0 -9.7 -1.8 4.3 4.1 -1.2 5.7 -1.9 -4.9 -3.3 -1.2
C13 Manufacture of textiles 1.6 1.1 4.3 2.5 -0.6 5.2 2.2 -4.8 -3.2 8.5 -1.8 -4.7 -1.3
C14 Manufacture of wearing apparel -1.2 -9.4 -4.3 -2.0 -6.5 3.3 5.5 -0.7 1.5 7.8 -4.4 -2.8 0.2
C15 Manufacture of leather and related products -3.8 -7.7 -5.1 -7.8 -5.1 -2.0 -1.1 -1.1 -2.4 2.7 1.4 -4.6 -0.8
Manufacture of wood and of products of wood and cork,
C16 except furniture; manufacture of articles of straw and -0.9 1.9 4.2 4.4 2.4 4.6 1.0 -6.3 -1.5 2.8 2.4 -2.3 -1.0
plaiting materials
C17 Manufacture of paper and paper products -0.2 3.8 4.4 4.9 2.0 4.8 3.9 0.9 -1.0 6.6 -1.3 0.1 1.0
C18 Printing and reproduction of recorded media -0.7 3.1 3.3 4.5 4.8 0.3 0.5 -0.2 -1.9 3.5 1.9 -1.6 0.3
C19 Manufacture of coke and refined petroleum products 1.0 4.9 3.5 5.0 0.7 2.9 0.0 -0.9 1.1 0.2 0.1 -2.9 -0.5
C20 Manufacture of chemicals and chemical products 0.5 4.2 2.7 5.5 5.4 4.4 4.6 -1.4 -7.4 12.0 0.5 -2.4 0.0
Manufacture of basic pharmaceutical products and
C21 10.4 6.5 5.2 0.5 6.4 5.9 -0.5 0.6 4.8 5.8 1.7 -1.6 2.3
pharmaceutical preparations
C22 Manufacture of rubber and plastic products 0.0 1.9 3.5 2.0 2.3 2.2 3.9 -4.2 -5.0 6.4 1.6 -3.7 -1.1
C23 Manufacture of other non-metallic mineral products 2.1 1.7 3.6 3.1 1.5 4.7 1.4 -4.1 -8.0 4.4 3.6 -5.1 -1.9
C24 Manufacture of basic metals 0.4 2.8 5.2 7.2 1.4 6.3 1.9 -2.8 -16.1 16.0 1.0 -3.1 -1.5
Manufacture of fabricated metal products, except machinery
C25 0.9 0.7 3.0 2.9 2.5 3.1 3.8 -5.7 -12.4 7.6 5.2 -3.7 -2.1
and equipment
C26 Manufacture of computer, electronic and optical products -6.0 -5.4 5.1 9.5 4.4 9.7 7.3 1.7 -5.7 9.9 5.2 -0.4 2.0
C27 Manufacture of electrical equipment 0.5 -1.3 2.3 3.9 2.9 5.8 2.6 -1.3 -8.8 7.7 1.0 -1.1 -0.6
C28 Manufacture of machinery and equipment n.e.c. 1.9 0.6 1.4 5.3 5.3 6.8 5.8 -0.3 -17.5 10.9 7.0 -0.2 -0.5
C29 Manufacture of motor vehicles, trailers and semi-trailers 1.2 2.5 2.6 4.1 1.7 3.9 5.3 -4.6 -12.7 16.8 7.3 -2.9 0.3
C30 Manufacture of other transport equipment 2.7 -1.6 2.4 2.8 2.4 6.7 3.8 2.5 -3.9 3.3 4.6 0.7 1.4
C31 Manufacture of furniture -3.0 -0.5 0.8 0.8 4.6 3.0 3.0 -2.2 -5.4 4.8 2.8 -2.9 -0.7
C32 Other manufacturing 3.4 6.4 0.3 0.9 3.8 6.2 1.0 -1.8 -1.3 8.0 0.7 -1.5 0.7
C33 Repair and installation of machinery and equipment 1.9 -1.1 1.4 7.4 1.1 6.9 3.9 2.1 -10.3 6.2 4.6 -2.1 -0.1
ELECTRICITY, GAS, STEAM AND AIR
D 3.8 6.1 8.0 4.3 2.3 2.4 0.3 0.0 -3.6 4.6 -5.2 3.1 -0.3
CONDITIONING SUPPLY
WATER SUPPLY; SEWERAGE, WASTE
E N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
MANAGEMENT AND REMEDIATION ACTIVITIES
F CONSTRUCTION 0.5 2.0 1.0 -0.3 -4.1 0.5 0.0 -1.6 2.0 2.7 1.0 -3.6 0.1
N/A: Data not available, Source: Eurostat

183
Table 6.6. EU-27 - Unit labour cost, annual growth rate (%)
Code Average
Sector 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
(NACE Rev. 2) 2007-2012

B MINING AND QUARRYING 7.1 -0.7 6.4 4.2 -0.4 7.4 4.8 10.7 11.7 1.6 10.2 9.1 8.6

C MANUFACTURING 3.0 1.7 0.3 -1.2 -0.5 -2.2 -0.1 6.0 10.6 -6.4 -0.6 4.5 2.7

C10 Manufacture of food products 2.1 0.9 2.1 -0.7 -0.7 0.4 1.4 4.9 1.2 -0.2 0.5 2.3 1.7
C11 Manufacture of beverages 1.3 -1.1 2.8 3.5 -1.2 -3.8 1.0 4.5 2.2 -0.7 -4.0 3.1 1.0
C12 Manufacture of tobacco products 5.2 2.2 5.8 19.1 6.8 2.2 -1.7 9.2 -3.1 0.9 1.3 4.2 2.4
C13 Manufacture of textiles 2.0 3.0 0.5 0.8 2.9 -2.3 0.4 9.6 6.3 -8.7 2.2 5.2 2.7
C14 Manufacture of wearing apparel 1.6 10.8 4.1 2.1 5.4 -0.9 2.1 8.2 4.9 -3.9 4.7 4.4 3.6
C15 Manufacture of leather and related products 9.8 8.2 4.1 8.1 5.7 5.8 9.2 10.9 5.5 -0.2 0.6 8.6 5.0
Manufacture of wood and of products of wood and cork,
C16 except furniture; manufacture of articles of straw and 5.2 -1.4 -1.9 -0.6 0.7 -0.3 5.0 12.0 5.5 -3.9 -0.4 1.6 2.8
plaiting materials
C17 Manufacture of paper and paper products 4.9 -2.5 -1.5 -1.5 0.7 -3.5 -1.3 3.6 3.8 -5.0 2.2 1.7 1.2
C18 Printing and reproduction of recorded media 4.8 0.3 -1.6 -1.5 -2.3 -0.5 0.4 4.2 2.4 -4.6 -2.5 3.0 0.4
C19 Manufacture of coke and refined petroleum products 1.7 4.8 -4.4 -0.5 2.2 3.8 2.2 5.1 6.4 3.5 3.2 4.0 4.4
C20 Manufacture of chemicals and chemical products 3.8 -1.1 1.7 -3.4 -0.9 -3.5 0.0 5.4 11.2 -9.1 4.5 3.7 2.9
Manufacture of basic pharmaceutical products and
C21 -6.1 -2.6 -0.1 1.6 -2.6 -2.7 5.7 -0.1 -2.5 -3.6 0.2 4.1 -0.4
pharmaceutical preparations
C22 Manufacture of rubber and plastic products 3.3 1.3 -0.1 0.6 0.1 -2.3 -0.6 7.7 8.3 -4.8 0.9 5.9 3.5
C23 Manufacture of other non-metallic mineral products 2.0 2.7 0.4 -0.8 0.5 -1.7 2.8 9.2 13.1 -3.0 -2.6 7.3 4.6
C24 Manufacture of basic metals -2.6 -0.8 -0.2 -3.6 3.1 -3.0 2.9 7.2 23.7 -13.8 1.4 6.4 4.3
Manufacture of fabricated metal products, except machinery
C25 4.2 1.8 -0.2 0.0 0.1 -0.9 0.9 10.5 16.1 -6.5 -2.6 5.9 4.3
and equipment
C26 Manufacture of computer, electronic and optical products 12.9 7.9 -4.7 -6.4 -2.6 -7.6 -4.1 1.7 11.3 -8.1 -2.7 4.4 1.1
C27 Manufacture of electrical equipment 3.4 3.3 -0.5 -0.5 -0.4 -4.2 1.2 5.8 12.6 -8.4 2.3 4.5 3.1
C28 Manufacture of machinery and equipment n.e.c. 3.1 2.5 1.7 -1.9 -2.5 -3.7 -1.5 4.3 27.9 -8.9 -3.6 3.9 4.0
C29 Manufacture of motor vehicles, trailers and semi-trailers 1.7 1.0 0.9 -2.1 0.2 -0.1 -5.2 9.2 17.6 -15.5 -3.5 7.3 2.4
C30 Manufacture of other transport equipment 3.4 7.7 1.5 -1.2 1.0 -4.1 0.1 2.4 8.1 2.1 -1.2 3.8 3.0
C31 Manufacture of furniture 6.0 4.7 -0.7 -0.9 -0.7 -0.8 0.3 7.4 10.5 -3.6 -2.5 4.4 3.1
C32 Other manufacturing 1.0 -1.2 3.3 1.1 -1.3 -3.0 4.2 3.7 4.4 -5.5 0.6 2.9 1.2
C33 Repair and installation of machinery and equipment 4.6 5.2 1.9 -2.7 1.6 -4.5 -0.2 3.8 14.0 -5.7 -3.9 4.1 2.2
ELECTRICITY, GAS, STEAM AND AIR
D 0.6 1.6 -1.7 -1.4 0.0 4.0 4.9 4.5 8.4 -1.7 6.1 1.5 3.7
CONDITIONING SUPPLY
WATER SUPPLY; SEWERAGE, WASTE
E N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
MANAGEMENT AND REMEDIATION ACTIVITIES

F CONSTRUCTION 7.2 5.8 2.6 3.5 9.3 3.0 6.5 6.7 0.4 -1.3 1.0 4.4 2.2

N/A: Data not available, Source: Eurostat

184
Table 6.7. EU-27 - Revealed comparative advantage index

Code
Sector 2007 2008 2009 2010 2011
(NACE Rev. 2)
C10 Manufacture of food products 1.20 1.12 1.10 1.09 1.06
C11 Manufacture of beverages 1.61 1.59 1.61 1.70 1.72
C12 Manufacture of tobacco products 1.52 1.56 1.61 1.67 1.72
C13 Manufacture of textiles 0.81 0.76 0.69 0.67 0.66
C14 Manufacture of wearing apparel 0.76 0.76 0.76 0.74 0.75
C15 Manufacture of leather and related products 0.96 0.91 0.91 0.87 0.91
Manufacture of wood and of products of wood and cork, except furniture; manufacture of articles
C16 1.15 1.18 1.18
of straw and plaiting materials 1.16 1.15
C17 Manufacture of paper and paper products 1.28 1.30 1.35 1.35 1.34
C18 Printing and reproduction of recorded media 1.20 1.62 1.79 1.88 1.87
C19 Manufacture of coke and refined petroleum products 0.84 0.84 0.77 0.79 0.78
C20 Manufacture of chemicals and chemical products 1.13 1.13 1.16 1.16 1.13
C21 Manufacture of basic pharmaceutical products and pharmaceutical preparations 1.47 1.53 1.54 1.65 1.62
C22 Manufacture of rubber and plastic products 1.18 1.21 1.18 1.19 1.19
C23 Manufacture of other non-metallic mineral products 1.22 1.19 1.18 1.15 1.13
C24 Manufacture of basic metals 0.92 0.88 0.82 0.85 0.86
C25 Manufacture of fabricated metal products, except machinery and equipment 1.18 1.19 1.16 1.20 1.20
C26 Manufacture of computer, electronic and optical products 0.60 0.60 0.57 0.57 0.58
C27 Manufacture of electrical equipment 0.98 0.99 0.98 0.97 0.99
C28 Manufacture of machineryand equipment n.e.c. 1.14 1.18 1.18 1.17 1.18
C29 Manufacture of motor vehicles, trailers and semi-trailers 1.22 1.22 1.30 1.28 1.32
C30 Manufacture of other transport equipment 0.85 0.88 1.15 1.21 1.15
C31 Manufacture of furniture 1.27 1.24 1.20 1.13 1.15
C32 Other manufacturing 0.80 0.78 0.75 0.77 0.72
Note: there was a transition from NACE REV 1 to NACE REV 2, therefore the data are only available from 2007
Source: own calculations using Comtrade data

185
Table 6.8. EU-27 - Relative trade balance (X-M)/(X+M)

Code
Sector 2007 2008 2009 2010 2011
(NACE Rev. 2)

C10 Manufacture of food products -0.03 -0.03 -0.02 -0.01 -0.01

C11 Manufacture of beverages 0.21 0.20 0.20 0.22 0.24
C12 Manufacture of tobacco products 0.03 0.06 0.06 0.05 0.08
C13 Manufacture of textiles 0.00 -0.01 -0.02 -0.02 -0.03
C14 Manufacture of wearing apparel -0.19 -0.19 -0.21 -0.22 -0.21
C15 Manufacture of leather and related products -0.07 -0.07 -0.08 -0.08 -0.06
Manufacture of wood and of products of wood and cork, except furniture; manufacture of articles
C16 0.00 0.02 0.04
of straw and plaiting materials 0.03 0.04
C17 Manufacture of paper and paper products 0.04 0.04 0.06 0.06 0.06
C18 Printing and reproduction of recorded media 0.08 0.05 0.04 0.08 0.09
C19 Manufacture of coke and refined petroleum products -0.03 -0.01 -0.05 -0.05 -0.05
C20 Manufacture of chemicals and chemical products 0.03 0.03 0.06 0.04 0.03
C21 Manufacture of basic pharmaceutical products and pharmaceutical preparations 0.07 0.08 0.08 0.10 0.10
C22 Manufacture of rubber and plastic products 0.04 0.04 0.04 0.05 0.05
C23 Manufacture of other non-metallic mineral products 0.08 0.08 0.09 0.08 0.08
C24 Manufacture of basic metals -0.06 -0.03 0.01 -0.01 0.00
C25 Manufacture of fabricated metal products, except machinery and equipment 0.09 0.09 0.10 0.10 0.10
C26 Manufacture of computer, electronic and optical products -0.11 -0.11 -0.11 -0.13 -0.10
C27 Manufacture of electrical equipment 0.07 0.08 0.08 0.07 0.07
C28 Manufacture of machineryand equipment n.e.c. 0.16 0.17 0.20 0.19 0.19
C29 Manufacture of motor vehicles, trailers and semi-trailers 0.06 0.08 0.08 0.12 0.13
C30 Manufacture of other transport equipment 0.13 0.11 0.11 0.10 0.14
C31 Manufacture of furniture 0.04 0.04 0.03 0.02 0.05
C32 Other manufacturing -0.04 -0.04 -0.04 -0.02 -0.02
Note: there was a transition from NACE REV 1 to NACE REV 2, therefore the data are only available from 2007
Source: own calculations using Comtrade data

186
Table 6.9.1. Revealed comparative advantage index in manufacturing industries in 2011 - EU countries, Japan and Brazil, China, India and Russia.
Non-
Leather Wood & Computers,
Refined Pharma- Rubber & metallic Basic Metal Electrical Motor Other Other
Food Bevarages Tobacco Textiles Clothing & wood Paper Printing Chemicals electronic Machinery Furniture
petroleum ceuticals plastics mineral metals products equipment vehicles transport manufacturing
footwear products & optical
products
C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32

Austria 0.87 2.24 0.28 0.68 0.55 0.70 4.49 2.19 1.30 0.24 0.47 1.52 1.30 1.36 1.34 2.16 0.43 1.35 1.40 1.34 0.70 1.19 0.71
Belgium 1.28 1.00 1.10 0.78 0.72 0.93 0.85 0.92 7.34 1.21 2.20 3.17 1.00 1.06 1.11 0.68 0.22 0.42 0.70 1.09 0.16 0.49 1.26
Bulgaria 1.31 0.81 4.94 1.14 2.87 1.18 1.63 0.76 0.22 1.55 0.55 0.90 0.92 2.20 2.83 0.86 0.27 1.11 0.93 0.36 0.24 1.31 0.35
Cyprus 2.16 1.22 27.96 0.13 0.32 0.82 0.20 0.34 0.00 0.00 0.97 7.44 0.30 0.34 0.78 0.68 0.72 0.71 0.41 0.27 1.15 0.47 1.20
Czech Rep. 0.44 0.57 1.57 0.88 0.31 0.46 1.42 0.94 1.74 0.20 0.53 0.32 1.67 1.63 0.64 2.14 1.11 1.66 1.16 2.00 0.39 1.52 0.73
Denmark 3.05 1.29 1.34 0.68 1.75 0.77 1.00 0.69 0.84 0.67 0.64 1.65 1.09 0.97 0.34 1.69 0.56 0.95 1.56 0.34 0.79 2.51 0.84
Estonia 1.00 2.10 0.20 1.17 1.04 0.86 7.71 0.83 0.73 2.33 0.64 0.10 1.29 1.38 0.49 1.79 0.95 1.40 0.75 0.61 0.24 2.74 0.53
Finland 0.35 0.49 0.04 0.26 0.19 0.25 5.28 9.87 0.74 1.60 0.85 0.62 0.92 0.77 1.66 1.04 0.47 1.32 1.46 0.27 0.37 0.23 0.46
France 1.18 4.63 0.59 0.54 0.70 1.18 0.63 1.03 1.80 0.51 1.30 1.70 1.10 0.99 0.75 0.90 0.48 0.88 0.87 1.15 3.97 0.52 0.76
Germany 0.74 0.65 2.05 0.53 0.50 0.39 0.81 1.20 2.49 0.21 1.00 1.34 1.29 1.02 0.76 1.31 0.58 1.22 1.60 1.91 1.30 0.85 0.57
Greece 2.16 1.52 4.89 1.08 1.47 0.55 0.55 0.58 1.14 4.56 0.71 1.26 0.98 1.33 1.93 0.84 0.22 0.65 0.29 0.09 0.50 0.31 0.32
Hungary 0.85 0.40 0.62 0.35 0.27 0.50 0.74 0.88 0.08 0.42 0.58 1.11 1.44 1.18 0.33 0.80 1.68 1.89 0.86 1.78 0.17 1.00 0.27
Ireland 1.44 1.76 0.53 0.10 0.15 0.09 0.41 0.11 0.00 0.23 2.85 9.43 0.32 0.22 0.08 0.28 0.65 0.24 0.31 0.03 0.41 0.09 1.49
Italy 0.87 2.30 0.03 1.35 1.58 3.09 0.53 1.03 0.98 0.70 0.70 1.10 1.35 1.90 1.09 1.68 0.23 1.05 1.82 0.73 0.75 2.38 0.95
Latvia 1.48 6.43 1.67 0.99 1.07 0.32 19.91 0.95 1.92 0.96 0.55 1.17 0.98 2.04 1.26 1.55 0.50 0.63 0.48 0.73 0.32 2.31 0.40
Lithuania 1.64 1.58 7.16 0.94 1.22 0.34 3.42 1.08 0.23 4.21 1.32 0.39 1.05 0.86 0.20 0.98 0.24 0.51 0.56 0.78 0.21 5.67 0.40
Luxembourg 0.89 1.02 6.08 2.17 0.32 0.45 2.24 1.79 0.01 0.01 0.49 0.15 4.16 2.51 3.86 1.20 0.28 0.74 0.74 0.63 1.09 0.12 0.21
Malta 0.53 0.31 0.69 1.05 0.10 0.12 0.01 0.02 0.93 4.67 0.13 2.04 1.12 0.33 0.05 0.23 1.87 0.92 0.24 0.03 1.64 0.08 1.78
Netherlands 1.94 1.29 5.35 0.45 0.61 0.66 0.31 0.86 0.29 2.02 1.62 0.95 0.79 0.49 0.61 0.79 1.05 0.54 1.07 0.39 0.32 0.40 0.80
Poland 1.46 0.45 4.79 0.60 0.71 0.41 2.33 1.66 0.54 0.59 0.76 0.32 1.85 1.61 0.92 1.79 0.60 1.35 0.57 1.64 1.29 5.03 0.27
Portugal 1.12 3.70 4.25 1.86 2.15 3.08 4.15 3.22 0.77 0.60 0.76 0.46 1.85 3.19 0.69 1.77 0.32 1.01 0.43 1.46 0.19 2.80 0.28
Romania 0.49 0.28 5.73 1.04 2.18 2.40 4.18 0.31 1.90 0.85 0.53 0.47 1.61 0.54 0.98 1.12 0.61 1.48 0.77 1.82 0.91 3.61 0.23
Slovakia 0.50 0.47 0.00 0.35 0.57 1.38 1.19 1.07 0.40 0.82 0.45 0.17 1.53 1.09 1.05 1.56 1.22 1.01 0.74 2.49 0.17 1.52 0.27
Slovenia 0.51 0.59 0.00 0.80 0.37 0.61 2.93 1.87 0.27 0.45 0.86 2.54 1.84 1.59 1.02 2.07 0.22 2.28 0.95 1.46 0.19 2.78 0.46
Spain 1.55 2.27 0.46 0.76 1.21 1.19 0.82 1.43 0.51 0.79 1.09 1.28 1.26 2.10 1.06 1.25 0.17 0.95 0.65 2.17 1.15 0.73 0.36
Sweden 0.49 0.84 0.28 0.29 0.36 0.21 3.54 5.49 0.22 1.06 0.67 1.37 0.85 0.60 1.11 1.11 0.82 1.01 1.25 1.34 0.31 1.55 0.43
United Kingdom 0.67 3.99 0.60 0.50 0.63 0.48 0.18 0.66 1.88 1.27 1.17 2.51 0.92 0.72 0.80 0.73 0.65 0.72 1.11 1.30 1.61 0.42 1.01

EU-27 1.06 1.72 1.72 0.66 0.75 0.91 1.15 1.34 1.87 0.78 1.13 1.62 1.19 1.13 0.86 1.20 0.58 0.99 1.18 1.32 1.15 1.15 0.72
USA 0.88 0.76 0.24 0.52 0.15 0.20 0.61 1.19 0.56 1.29 1.41 0.99 0.97 0.73 0.72 0.89 1.00 0.86 1.36 1.04 0.41 0.48 1.52
Japan 0.07 0.06 0.08 0.43 0.02 0.02 0.02 0.26 0.18 0.32 0.94 0.17 1.09 1.04 1.11 0.73 1.08 1.09 2.09 2.01 1.35 0.14 0.45
Brazil 5.17 0.11 0.47 0.37 0.04 1.74 1.73 2.99 0.34 0.38 0.94 0.39 0.72 0.98 1.75 0.73 0.10 0.43 0.82 1.04 1.42 0.52 0.17
China 0.37 0.09 0.15 2.54 2.72 2.52 0.93 0.43 0.23 0.21 0.53 0.23 1.00 1.53 0.53 1.34 1.88 1.47 0.74 0.28 0.86 2.12 1.29
India 1.35 0.10 0.47 2.82 1.94 1.18 0.11 0.23 0.74 3.07 0.93 1.02 0.61 0.74 0.77 0.94 0.19 0.38 0.39 0.32 1.23 0.32 5.37
Russia 0.49 0.22 1.07 0.05 0.02 0.11 3.45 0.99 0.15 7.83 1.50 0.05 0.22 0.35 2.61 0.27 0.10 0.17 0.13 0.10 0.46 0.12 0.03
Source: Own calculations using COMTRADE data

187
Table 6.9.2. Relative trade balance (X-M)/(X+M) in manufacturing industries in 2011 - EU countries, Japan and Brazil, China, India and Russia.
Non-
Wood & Computers, Other
Leather & Refined Pharma- Rubber & metallic Basic Metal Electrical Motor Other
Food Bevarages Tobacco Textiles Clothing wood Paper Printing Chemicals electronic Machinery Furniture manufac-
footwear petroleum ceuticals plastics mineral metals products equipment vehicles transport
products & optical turing
products
C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32

Austria -0.06 0.53 -0.71 -0.03 -0.42 -0.20 0.42 0.23 -0.37 -0.55 -0.27 0.03 -0.04 -0.01 0.03 0.11 -0.09 0.11 0.06 0.02 0.15 -0.21 -0.10
Belgium 0.12 -0.02 0.08 0.23 -0.01 0.22 -0.01 -0.05 0.04 0.02 0.13 0.13 0.04 0.11 0.18 -0.04 -0.19 -0.06 0.05 -0.04 -0.12 -0.21 0.03
Bulgaria -0.12 -0.19 0.49 -0.43 0.56 0.09 0.25 -0.34 -0.75 0.07 -0.34 -0.21 -0.23 0.21 0.40 -0.14 -0.38 -0.01 -0.04 -0.27 -0.23 0.27 -0.16
Cyprus -0.67 -0.87 -0.32 -0.92 -0.93 -0.82 -0.96 -0.92 -1.00 -1.00 -0.62 0.03 -0.91 -0.95 -0.55 -0.80 -0.54 -0.78 -0.77 -0.89 -0.52 -0.93 -0.61
Czech Rep. -0.20 0.02 0.32 0.09 -0.20 -0.17 0.34 -0.06 0.01 -0.22 -0.20 -0.39 0.01 0.21 -0.20 0.18 -0.04 0.15 0.16 0.34 0.34 0.28 0.18
Denmark 0.25 -0.16 0.29 -0.03 -0.02 -0.20 -0.37 -0.36 -0.24 -0.22 -0.09 0.21 -0.08 -0.12 -0.34 0.13 -0.12 -0.06 0.26 -0.38 0.06 0.21 -0.04
Estonia -0.04 -0.25 -0.73 0.06 0.12 0.05 0.44 -0.13 -0.55 -0.11 -0.19 -0.72 -0.11 0.04 -0.28 0.21 0.01 0.04 -0.07 -0.10 -0.32 0.61 0.04
Finland -0.44 -0.39 -0.95 -0.37 -0.67 -0.43 0.56 0.80 -0.54 0.21 -0.05 -0.18 0.01 -0.12 0.32 0.03 -0.18 0.12 0.20 -0.50 0.08 -0.61 -0.10
France -0.05 0.62 -0.60 -0.15 -0.38 -0.10 -0.34 -0.19 0.20 -0.37 0.02 0.04 -0.12 -0.17 -0.08 -0.15 -0.20 -0.06 -0.05 -0.06 0.25 -0.51 -0.16
Germany 0.05 -0.06 0.51 0.02 -0.32 -0.27 0.06 0.14 0.40 -0.38 0.12 0.13 0.20 0.18 0.00 0.23 -0.02 0.21 0.39 0.39 0.07 -0.08 0.05
Greece -0.31 -0.24 0.08 -0.15 -0.31 -0.65 -0.50 -0.68 -0.49 0.25 -0.50 -0.57 -0.25 -0.24 0.17 -0.24 -0.62 -0.35 -0.52 -0.79 -0.69 -0.73 -0.63
Hungary 0.15 0.04 -0.14 -0.18 -0.06 -0.09 0.12 -0.04 -0.83 0.04 -0.09 0.04 0.05 0.15 -0.31 -0.14 0.14 0.15 -0.07 0.41 0.18 0.42 0.05
Ireland 0.28 0.22 0.05 -0.32 -0.61 -0.59 0.06 -0.69 -0.95 -0.49 0.67 0.76 -0.23 -0.29 -0.26 -0.05 0.28 -0.10 0.15 -0.75 -0.46 -0.55 0.55
Italy -0.13 0.62 -0.98 0.17 0.12 0.27 -0.41 -0.05 -0.04 0.24 -0.22 -0.12 0.24 0.42 -0.06 0.42 -0.41 0.20 0.47 -0.12 0.22 0.64 0.14
Latvia -0.23 0.31 -0.30 -0.11 -0.07 -0.48 0.79 -0.37 -0.45 -0.50 -0.32 -0.19 -0.30 0.02 -0.04 -0.04 -0.13 -0.24 -0.40 -0.22 -0.30 0.33 -0.35
Lithuania 0.11 -0.15 0.54 -0.11 0.26 -0.22 0.26 -0.15 -0.70 0.75 0.04 -0.39 0.01 -0.15 -0.42 0.05 -0.23 -0.08 -0.12 -0.10 -0.35 0.82 0.04
Luxembourg -0.29 -0.54 -0.06 0.61 -0.53 -0.32 0.14 -0.07 -0.98 -0.99 -0.46 -0.68 0.27 0.04 0.32 -0.15 -0.31 -0.12 -0.01 -0.48 -0.50 -0.89 -0.47
Malta -0.60 -0.77 -0.63 0.41 -0.80 -0.73 -0.98 -0.98 -0.54 -0.39 -0.79 0.26 -0.07 -0.72 -0.78 -0.63 0.10 -0.01 -0.46 -0.88 -0.48 -0.91 0.43
Netherlands 0.24 0.15 0.62 0.07 -0.12 -0.05 -0.51 -0.05 -0.30 0.12 0.21 0.08 -0.01 -0.17 -0.03 0.03 0.03 -0.04 0.23 -0.15 0.11 -0.31 0.01
Poland 0.17 -0.09 0.78 -0.30 -0.06 -0.32 0.39 -0.03 -0.30 0.07 -0.24 -0.47 0.07 0.15 -0.05 0.10 -0.16 0.12 -0.24 0.25 0.23 0.75 -0.24
Portugal -0.35 0.43 0.47 0.06 0.15 0.22 0.39 0.27 -0.11 -0.20 -0.34 -0.54 0.12 0.38 -0.19 0.18 -0.32 -0.04 -0.27 -0.05 -0.15 0.30 -0.48
Romania -0.40 -0.45 0.68 -0.47 0.53 0.08 0.59 -0.64 -0.14 0.09 -0.38 -0.52 -0.18 -0.47 -0.09 -0.28 -0.19 -0.11 -0.25 0.25 0.57 0.63 -0.27
Slovakia -0.10 -0.11 -1.00 -0.16 0.02 0.23 0.28 0.16 -0.05 0.40 -0.08 -0.53 0.16 0.11 0.23 0.06 0.12 0.03 0.08 0.32 0.10 0.26 0.01
Slovenia -0.35 -0.14 -1.00 0.03 -0.36 -0.34 0.17 0.09 -0.63 -0.61 -0.17 0.41 0.15 0.03 -0.12 0.20 -0.24 0.33 0.12 0.05 -0.01 0.33 -0.07
Spain 0.03 0.25 -0.75 -0.05 -0.28 -0.11 -0.02 0.02 -0.42 -0.22 -0.11 -0.11 0.04 0.37 0.13 0.11 -0.57 0.01 -0.07 0.16 0.38 -0.22 -0.38
Sweden -0.32 -0.13 -0.40 -0.25 -0.41 -0.47 0.51 0.71 -0.62 0.15 -0.13 0.32 -0.07 -0.23 0.13 0.09 -0.01 0.02 0.14 0.11 -0.09 0.09 -0.09
United Kingdom -0.44 0.14 -0.47 -0.31 -0.57 -0.55 -0.81 -0.47 0.44 0.09 -0.06 0.13 -0.21 -0.27 -0.08 -0.24 -0.24 -0.21 0.01 -0.14 0.16 -0.63 -0.16

EU-27 -0.01 0.24 0.08 -0.03 -0.21 -0.06 0.04 0.06 0.09 -0.05 0.03 0.10 0.05 0.08 0.00 0.10 -0.10 0.07 0.19 0.13 0.14 0.05 -0.02
USA -0.04 -0.48 -0.18 -0.33 -0.88 -0.82 -0.39 0.04 0.48 0.00 0.13 -0.26 -0.17 -0.25 -0.20 -0.15 -0.26 -0.25 -0.03 -0.30 -0.46 -0.69 -0.19
Japan -0.87 -0.83 -0.96 -0.18 -0.97 -0.95 -0.98 -0.29 0.38 -0.46 0.15 -0.64 0.35 0.28 0.30 0.18 0.12 0.30 0.63 0.76 0.60 -0.66 -0.26
Brazil 0.78 -0.81 0.91 -0.51 -0.81 0.54 0.82 0.54 -0.29 -0.69 -0.46 -0.59 -0.29 -0.08 0.33 -0.22 -0.85 -0.45 -0.38 -0.22 0.02 0.30 -0.53
China 0.07 -0.43 0.66 0.72 0.95 0.81 0.26 -0.12 0.40 -0.20 -0.22 0.14 0.42 0.61 -0.03 0.62 0.24 0.35 -0.02 -0.15 0.41 0.91 0.73
India 0.32 -0.15 0.83 0.65 0.95 0.65 -0.47 -0.54 -0.31 0.56 -0.30 0.51 0.14 0.03 -0.65 0.15 -0.61 -0.29 -0.44 0.15 0.24 0.03 0.38
Russia -0.61 -0.81 0.31 -0.89 -0.97 -0.89 0.61 -0.20 -0.89 0.91 0.15 -0.95 -0.75 -0.65 0.49 -0.73 -0.82 -0.82 -0.88 -0.91 -0.51 -0.83 -0.92
Source: Own calculations using COMTRADE data

188
Table 6.10. Revealed comparative advantage index in service industries in 2011 - EU countries, US, Japan, Brazil, China, India and Russia.
Telecom.,
Insurance and Other business Personal, cultural
Country name computer and Construction Finance Transport Travel
pension services and recreational
information
Austria 0.56 0.52 0.41 0.81 1.10 0.43 1.17 1.25
Belgium 0.86 1.17 0.70 0.65 1.62 0.69 1.29 0.45
Bulgaria 0.89 0.42 0.20 0.98 0.52 0.74 0.97 2.00
Cyprus 0.13 0.27 2.20 0.28 1.14 0.49 1.23 1.22
Czech Republic 0.90 1.44 0.07 0.64 1.23 0.89 1.08 1.21
Denmark 0.36 0.21 0.21 0.21 0.88 0.72 2.78 0.37
Estonia 0.76 2.28 0.27 0.07 0.78 0.29 1.73 0.81
Finland 2.51 1.17 0.50 0.10 0.78 0.30 0.65 0.54
France 0.47 1.81 0.59 1.32 1.23 1.67 1.00 0.97
Germany 0.78 1.85 1.04 1.17 1.19 0.33 1.08 0.53
Greece 0.23 1.28 0.08 0.72 0.26 0.46 2.26 1.38
Hungary 0.66 0.76 0.16 0.08 1.08 5.64 0.96 0.92
Ireland 4.06 0.00 1.67 5.63 0.93 0.00 0.26 0.16
Italy 0.79 0.05 0.47 1.16 1.29 0.25 0.67 1.52
Latvia 0.53 0.75 1.29 0.28 0.71 0.40 2.26 0.62
Lithuania 0.26 0.87 0.17 -0.01 0.34 0.32 2.56 0.90
Luxembourg 0.45 0.22 11.17 2.30 0.69 3.34 0.24 0.25
Malta 0.28 0.00 1.02 0.35 0.49 35.88 0.39 0.95
Netherlands 1.07 1.01 0.26 0.34 1.66 0.62 1.28 0.50
Poland 0.68 1.70 0.24 0.57 1.28 1.18 1.35 1.06
Portugal 0.43 1.23 0.23 0.30 0.68 1.10 1.30 1.65
Romania 1.35 1.53 0.30 0.56 1.00 0.86 1.20 0.44
Slovak Republic 0.92 1.49 0.10 0.20 0.75 1.09 1.48 1.37
Slovenia 0.74 1.34 0.13 0.88 0.70 0.91 1.29 1.55
Spain 0.61 1.23 0.69 0.51 1.14 1.28 0.77 1.56
Sweden 1.51 0.54 0.44 0.72 1.39 0.75 0.81 0.78
United Kingdom 0.76 0.35 4.16 3.16 1.52 1.14 0.58 0.45

EU27 1.11 0.75 1.28 1.19 1.13 1.10 1.06 0.88

United States 0.44 0.20 2.34 1.33 0.76 0.40 0.61 0.92
Japan 0.13 3.20 0.57 0.62 1.29 0.10 1.29 0.30
Brazil 0.13 0.02 1.33 0.69 2.50 0.12 0.71 0.64
China 0.71 3.19 0.09 0.85 1.23 0.06 0.89 0.98
India 4.33 0.24 0.87 0.98 0.77 4.52 0.59 0.48
Russian Federation 0.51 3.04 0.36 0.30 1.23 0.72 1.38 0.73
Source: Own calculations based on IMF and OECD data

189
TOWARDS
KNOWLEDGE DRIVEN
REINDUSTRIALISATION
European Competitiveness Report 2013